Showing 1–50 of 186 results for author: Teng, F

One Score, Two Decisions: Selective Prediction on the Rare-Disease Tail

Authors: Zhaoyang Jiang, Zhizhong Fu, Yunsoo Kim, Zicheng Li, Xuanqi Peng, Fei Teng, Jiacong Mi, Honghan Wu

Abstract: Given a patient's clinical findings, a diagnostic system ranks possible diseases and must decide when to endorse its first prediction or defer it for review. This decision is usually made by thresholding the top score. Selective prediction over ranked outputs begins with two checks. First, the ranker must produce enough correct top-ranked predictions to make the target feasible. Across 2,000 patie… ▽ More Given a patient's clinical findings, a diagnostic system ranks possible diseases and must decide when to endorse its first prediction or defer it for review. This decision is usually made by thresholding the top score. Selective prediction over ranked outputs begins with two checks. First, the ranker must produce enough correct top-ranked predictions to make the target feasible. Across 2,000 patient records stratified by disease prevalence, eight small open-weight LLMs achieve at most 4.6% Recall@1 on ultra-rare diseases. At 10% coverage, even a perfect confidence ranking of their existing predictions therefore cannot reach 50% selective accuracy. More accurate models pass the same check, showing that the limit is regime-specific. Second, the confidence signal must match the decision being made. For fixed-candidate rankers, the top-two margin cancels components shared across candidates. On phenotype-only Exomiser, it selects 10% of cases at 29.0% accuracy, compared with 13.3% overall, while the top score provides no reliable gate. Yet that cancellation can remove information needed to detect whether the candidate list contains an answer. SciFact retrieval and biomedical entity linking confirm this distinction. Finally, we prove that unlabelled scores alone cannot determine whether switching to the margin will help. △ Less

Submitted 4 August, 2026; originally announced August 2026.

ToolArtist: Tool-Using Unified Multimodal Models for Agentic Image Generation

Authors: Jiahao Zhao, Xiaomin Yu, Zhongxiang Sun, Fengwei Teng, Chengwei Qin, Xiaobin Hu, Jun Xu, Shuicheng Yan

Abstract: Text-to-image (T2I) models can produce visually compelling images, yet they remain limited on open-world tasks that require complex semantic understanding, multi-step reasoning, and the integration of external world knowledge. Existing efforts introduce agent capabilities into image generation, but they either prescribe a fixed workflow or place only a subset of the open-world image generation pro… ▽ More Text-to-image (T2I) models can produce visually compelling images, yet they remain limited on open-world tasks that require complex semantic understanding, multi-step reasoning, and the integration of external world knowledge. Existing efforts introduce agent capabilities into image generation, but they either prescribe a fixed workflow or place only a subset of the open-world image generation process under agent control. Consequently, reasoning, tool invocation, and image generation are not coordinated by a single policy. We propose ToolArtist, a fully agentic image generation model obtained by post-training a Unified Multimodal Model (UMM). ToolArtist dynamically orchestrates reasoning, external tool use, and native image generation within one unified policy. During Supervised Fine-Tuning (SFT), we equip a teacher agent with search tools alongside an image-generation tool. We then convert the collected trajectories into a UMM compatible format, where the image-generation tool is concealed while the resulting generated images are retained. During Reinforcement Learning (RL), we develop an agentic RL infrastructure for UMMs and introduce Reason-Act-Draw GRPO (RAD-GRPO), which uses complementary intent and quality rewards to jointly optimize the model. Experiments show that placing the entire open-world image-generation process under an agent policy consistently outperforms approaches with fixed pipelines or only partially agent-controlled components. We release the training data and the complete post-training infrastructure. △ Less

Submitted 5 August, 2026; originally announced August 2026.

Structured Differentiable Optimization for Efficient Decision-focused Learning in Power Systems

Authors: Wangkun Xu, Fei Teng

Abstract: Decision-focused learning (DfL) trains forecasting models to align downstream decision consequences, such as power-system operating costs. However, its application to realistic power networks is limited by the need to repeatedly solve and differentiate large optimization problems during training. This paper presents DiffAPQP, a solver-flexible framework and open-source Python package for scalable… ▽ More Decision-focused learning (DfL) trains forecasting models to align downstream decision consequences, such as power-system operating costs. However, its application to realistic power networks is limited by the need to repeatedly solve and differentiate large optimization problems during training. This paper presents DiffAPQP, a solver-flexible framework and open-source Python package for scalable DfL with affine-parametric quadratic programs. To accelerate the forward pass, DiffAPQP automatically canonicalizes quadratic power-system models written in CVXPY into a differentiation-ready representation and takes advantage of the repetitive solving structure through solver warm-start and solver-data update during training. For the backward pass acceleration, we establish the equivalence between differentiation through the full KKT system and a reduced system obtained by eliminating inactive inequality constraints. For training losses depending solely on the optimal value, we further derive an envelope-theorem-based gradient that avoids solving an adjoint KKT system, resulting in eligible backward time. To our knowledge, this work presents the first solver-based end-to-end DfL demonstration on the IEEE 118-bus system with a 24-hour coupled economic-dispatch and redispatch horizon. Under matched SCS and Clarabel backends on a Linux machine, DiffAPQP achieves $2.27\times$--$3.58\times$ closed-loop and $3.62\times$--$4.38\times$ counterfactual end-to-end DfL training speedups over CvxpyLayers. The best solver configurations increase these speedups to $3.91\times$ (from 38.65 to 9.55 min/epoch) and $6.39\times$ (from 10.73 to 1.68 min/epoch), respectively. Additionally, DiffAPQP reduces peak memory usage by approximately $50\%$, while keeping similar operating costs as CvxpyLayers. △ Less

Submitted 4 August, 2026; originally announced August 2026.

Presentation, Not Mechanism: A Render Confound in Deprecation-Aware Memory Evaluation

Authors: Zhaoyang Jiang, Zhizhong Fu, Zicheng Li, Yunsoo Kim, Jiacong Mi, Xuanqi Peng, Fei Teng, Honghan Wu

Abstract: AI systems increasingly retrieve from records that revise themselves: issue threads, encyclopedic histories, policy logs, and long conversations. The challenge is not only finding relevant evidence, but deciding which claims remain in force, which were superseded, and when to abstain. Structured memories promise to solve this with typed edges, temporal updates, and conflict status, yet evaluations… ▽ More AI systems increasingly retrieve from records that revise themselves: issue threads, encyclopedic histories, policy logs, and long conversations. The challenge is not only finding relevant evidence, but deciding which claims remain in force, which were superseded, and when to abstain. Structured memories promise to solve this with typed edges, temporal updates, and conflict status, yet evaluations often change mechanism and prompt presentation together. We study this as Evidence-State Revision, comparing flat retrieval, coarse edge invalidation, and fine-grained RevisionLedger on 2,907 high-agreement questions from GitHub, multi-repo issue histories, Wikipedia, and DyKnow-style temporal streams. A render-matched control (same layout, deprecation disabled) reveals the central confound: when a value is changed and later restored, RevisionLedger appears to beat a flat baseline by +0.182, but almost all the gain comes from easier presentation; the fine-grained mechanism residual is indistinguishable from zero (+0.021 to +0.025 across two judge families). After presentation is controlled, coarse invalidation is the only mechanism that pays for current-state queries, beating the fine ledger by 0.084; the same query-sufficiency principle says provenance mainly needs retained invalidated evidence, not richer typing. Memory evaluations should hold render fixed, and deprecation-aware systems should deploy the coarsest retained state that covers their queries. △ Less

Submitted 17 July, 2026; originally announced July 2026.

PS-MOT: Cultivating Instance Awareness from Point Seeds for Multi-Object Tracking

Authors: Kai Luo, Fei Teng, Mengfei Duan, Wanjun Jia, Xu Wang, Hao Shi, Kunyu Peng, Zhiyong Li, Kailun Yang

Abstract: We introduce Point-supervised Multi-Object Tracking (PS-MOT) as a cost-effective alternative to traditional bounding box supervision, shifting the focus from spatial fitting to topological center-driven representation. However, PS-MOT faces challenges, e.g., spatial ambiguity and identity drift due to the lack of explicit geometric structure and scale constraints. To address these, we propose PS-T… ▽ More We introduce Point-supervised Multi-Object Tracking (PS-MOT) as a cost-effective alternative to traditional bounding box supervision, shifting the focus from spatial fitting to topological center-driven representation. However, PS-MOT faces challenges, e.g., spatial ambiguity and identity drift due to the lack of explicit geometric structure and scale constraints. To address these, we propose PS-Track, a hierarchical pipeline transitioning from points to instances across data, model, and loss levels. At the data level, we introduce Temporal-Feedback Prompting (TFP) to evolve points into temporally consistent pseudo-labels using negative spatial cues and motion priors. At the model level, we design the Point-Excited Wavelet Attention (PEWA) module, which leverages semantic correlations to activate high-frequency components, ``hallucinating'' object boundaries. At the loss level, Uncertainty-Guided Gaussian Learning (UGL) models pseudo-labels as probabilistic distributions, dynamically calibrating supervision intensity. Experiments on DanceTrack, EmboTrack, SportsMOT, and JRDB demonstrate that PS-Track provides a feasible and effective point-supervised alternative across diverse tracking scenarios, establishing a new state-of-the-art for point-supervised tracking. The source code is available at https://github.com/xifen523/PS-MOT. △ Less

Submitted 29 June, 2026; originally announced June 2026.

Comments: Accepted to ECCV 2026. The source code is available at https://github.com/xifen523/PS-MOT

Input Convex Neural Network as a Surrogate in Stability-Constrained Optimization for IBR-dominated Power Systems

Authors: Wangkun Xu, Hongyang Jia, Yi Wang, Ning Zhang, Fei Teng

Abstract: Input convex neural networks (ICNNs) are increasingly used as surrogates for stability indices and embedded as constraints in power-system optimization. This letter clarifies two recurring formulation limitations that can negate ICNN convexity benefits: (i) applying generic Big-$M$ mixed-integer reformulations introduces auxiliary binaries that are unnecessary for enforcing ICNN sublevel constrain… ▽ More Input convex neural networks (ICNNs) are increasingly used as surrogates for stability indices and embedded as constraints in power-system optimization. This letter clarifies two recurring formulation limitations that can negate ICNN convexity benefits: (i) applying generic Big-$M$ mixed-integer reformulations introduces auxiliary binaries that are unnecessary for enforcing ICNN sublevel constraints; and (ii) reversing the stability inequality transforms a convex sublevel set into a generally nonconvex superlevel set, invalidating global-convergence guarantees of cut-based methods. After clarifying the limitations, we provide (i) an exact LP-based epigraph reformulation for ReLU-ICNNs, (ii) an outer-approximation scheme with global guarantees under the sublevel convention, and (iii) a feasibility-preserving inner-approximation scheme for the superlevel convention, with simulations on IEEE 14- and 118-bus unit commitment instances. △ Less

Submitted 24 June, 2026; originally announced June 2026.

From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Long-Context LLMs

Authors: Zhanchao Xu, Haoyang Li, Qingfa Xiao, Fei Teng, Chen Jason Zhang, Lei Chen, Qing Li

Abstract: Existing sparse attention and KV cache compression methods for long-context LLM inference typically apply fixed sparsity patterns or uniform budgets across all attention heads, overlooking the substantial variation in attention behavior among heads and contexts. We observe two distinct entropy patterns among attention heads: Rigid Heads, whose entropy stays near zero across input segments, and Dyn… ▽ More Existing sparse attention and KV cache compression methods for long-context LLM inference typically apply fixed sparsity patterns or uniform budgets across all attention heads, overlooking the substantial variation in attention behavior among heads and contexts. We observe two distinct entropy patterns among attention heads: Rigid Heads, whose entropy stays near zero across input segments, and Dynamic Heads, whose entropy fluctuates significantly. Crucially, the distribution of these types is context-dependent and cannot be predetermined offline. We therefore propose EntropyInfer, a training-free framework that uses attention entropy to adaptively allocate compute at the granularity of individual heads and segments during prefilling. For decoding, we introduce a latent KV cache compression scheme that leverages generated output tokens, rather than prefill tokens alone, to identify and retain the most critical cache entries. Extensive experiments on Llama, Qwen and openPangu model series show that EntropyInfer consistently outperforms baselines including SnapKV, AdaKV, and CritiPrefill, achieving up to 2.39$\times$ end-to-end speedup beyond 100k tokens with minimal quality degradation compared to full attention. The code is released in https://github.com/SHA-4096/EntropyInfer. △ Less

Submitted 8 June, 2026; originally announced June 2026.

Better Accuracies, Worse Reasoning: A Step-Level Audit of Medical Chain-of-Thought Distillation

Authors: Zhaoyang Jiang, Xuanqi Peng, Fei Teng, Zhizhong Fu, Yunsoo Kim, Jiacong Mi, Zicheng Li, Honghan Wu

Abstract: Chain-of-thought (CoT) distillation trains a smaller model to imitate a teacher's reasoning trace, but it is typically evaluated by final-answer metrics including accuracy. We ask whether gains in answer quality are accompanied by improvements in the trace. In medical QA, where short answer options can leave a richer clinical justification under-specified, a Qwen3-8B student distilled from a DeepS… ▽ More Chain-of-thought (CoT) distillation trains a smaller model to imitate a teacher's reasoning trace, but it is typically evaluated by final-answer metrics including accuracy. We ask whether gains in answer quality are accompanied by improvements in the trace. In medical QA, where short answer options can leave a richer clinical justification under-specified, a Qwen3-8B student distilled from a DeepSeek-V3-family teacher improves on MedQA-USMLE answer metrics (SC@64 74.7% to 84.4%; expected calibration error (ECE) 0.096 to 0.034). Yet under a Kimi-K2.6 style-blind LLM-judge audit, its error rate over non-abstained steps rises from 30.6% to 50.3%. In this primary medical setting, answer quality and trace factuality move in opposite directions. This before--after pattern persists across evaluators, teacher strengths, student scales and families, medical benchmarks, and style, segmentation, and answer-correctness controls. A 150-step blinded audit by a clinical expert reproduces the same ordering. Boundary checks narrow the scope of the claim: the risk appears when a compact answer under-constrains the rationale and a capable student can imitate expert-like form without reliably grounding each local claim. Standard answer metrics and aggregate hedging rates do not reveal the shift. When such traces are released or reused, answer-level metrics alone are insufficient. △ Less

Submitted 27 May, 2026; originally announced May 2026.

Distributed Synthesis of Gray-Box Distributed H2 Controllers

Authors: Michael C. A. Nestor, Fei Teng

Abstract: Distributed controller synthesis offers scalable and privacy-preserving control design, but typical state-of-the-art approaches either assume white-box models or resort to centralized synthesis. In this paper, we combine partially known model knowledge and an input-state dataset within a distributed gray-box scheme to design \(\mathcal{H}_2\) controllers. Our method can handle unknown dynamics and… ▽ More Distributed controller synthesis offers scalable and privacy-preserving control design, but typical state-of-the-art approaches either assume white-box models or resort to centralized synthesis. In this paper, we combine partially known model knowledge and an input-state dataset within a distributed gray-box scheme to design \(\mathcal{H}_2\) controllers. Our method can handle unknown dynamics and offers scalable synthesis. Each agent communicates with a set of neighbors determined by the physical coupling topology of the system such that we can apply the Alternating Direction Method of Multipliers (ADMM) to solve the problem iteratively in a fully distributed fashion (i.e., without a central server). The effectiveness and flexibility of the proposed approach is demonstrated in simulations of the IEEE 39-bus power system test case. △ Less

Submitted 17 May, 2026; originally announced May 2026.

Comments: Accepted for presentation at the 23rd IFAC World Congress, Busan, South Korea, 2026

LeHome: A Simulation Environment for Deformable Object Manipulation in Household Scenarios

Authors: Zeyi Li, Yushi Yang, Shawn Xie, Kyle Xu, Tianxing Chen, Yuran Wang, Zhenhao Shen, Yan Shen, Yue Chen, Wenjun Li, Yukun Zheng, Chaorui Zhang, Siyi Lin, Fei Teng, Hongjun Yang, Ming Chen, Steve Xie, Ruihai Wu

Abstract: Household environments present one of the most common, impactful yet challenging application domains for robotics. Within household scenarios, manipulating deformable objects is particularly difficult, both in simulation and real-world execution, due to varied categories and shapes, complex dynamics, and diverse material properties, as well as the lack of reliable deformable-object support in exis… ▽ More Household environments present one of the most common, impactful yet challenging application domains for robotics. Within household scenarios, manipulating deformable objects is particularly difficult, both in simulation and real-world execution, due to varied categories and shapes, complex dynamics, and diverse material properties, as well as the lack of reliable deformable-object support in existing simulations. We introduce LeHome, a comprehensive simulation environment designed for deformable object manipulation in household scenarios. LeHome covers a wide spectrum of deformable objects, such as garments and food items, offering high-fidelity dynamics and realistic interactions that existing simulators struggle to simulate accurately. Moreover, LeHome supports multiple robotic embodiments and emphasizes low-cost robots as a core focus, enabling end-to-end evaluation of household tasks on resource-constrained hardware. By bridging the gap between realistic deformable object simulation and practical robotic platforms, LeHome provides a scalable testbed for advancing household robotics. Webpage: https://lehome-web.github.io/ . △ Less

Submitted 24 April, 2026; originally announced April 2026.

Comments: ICRA2026 Accepted

IAT: Instance-As-Token Compression for Historical User Sequence Modeling in Industrial Recommender Systems

Authors: Xinchun Li, Ning Zhang, Qianqian Yang, Fei Teng, Wenlin Zhao, Huizhi Yang, Heng Shi, Linlan Chen, Yixin Wu, Zhen Wang, Daiye Hou, Fei Qin, Lele Yu, Yaocheng Tan

Abstract: Although sophisticated sequence modeling paradigms have achieved remarkable success in recommender systems, the information capacity of hand-crafted sequential features constrains the performance upper bound. To better enhance user experience by encoding historical interaction patterns, this paper presents a novel two-stage sequence modeling framework termed Instance-As-Token (IAT). The first stag… ▽ More Although sophisticated sequence modeling paradigms have achieved remarkable success in recommender systems, the information capacity of hand-crafted sequential features constrains the performance upper bound. To better enhance user experience by encoding historical interaction patterns, this paper presents a novel two-stage sequence modeling framework termed Instance-As-Token (IAT). The first stage of IAT compresses all features of each historical interaction instance into a unified instance embedding, which encodes the interaction characteristics in a compact yet informative token. Both temporal-order and user-order compression schemes are proposed, with the latter better aligning with the demands of downstream sequence modeling. The second stage involves the downstream task fetching fixed-length compressed instance tokens via timestamps and adopting standard sequence modeling approaches to learn long-range preferences patterns. Extensive experiments demonstrate that IAT significantly outperforms state-of-the-art methods and exhibits superior in-domain and cross-domain transferability. IAT has been successfully deployed in real-world industrial recommender systems, including e-commerce advertising, shopping mall marketing, and live-streaming e-commerce, delivering substantial improvements in key business metrics. △ Less

Submitted 10 April, 2026; originally announced April 2026.

Skip-Connected Policy Optimization for Implicit Advantage

Authors: Fengwei Teng, Jinyi Bai, Xinhao Yao, Demi Ruohan Wang, Jiahao Zhao, Zhijiang Guo

Abstract: Group Relative Policy Optimization (GRPO) has proven effective in RLVR by using outcome-based rewards. While fine-grained dense rewards can theoretically improve performance, we reveal that under practical sampling budgets, Monte Carlo estimation yields high-variance and sign-inconsistent advantages for early reasoning tokens, paradoxically underperforming outcome-only GRPO. We propose Skip-Connec… ▽ More Group Relative Policy Optimization (GRPO) has proven effective in RLVR by using outcome-based rewards. While fine-grained dense rewards can theoretically improve performance, we reveal that under practical sampling budgets, Monte Carlo estimation yields high-variance and sign-inconsistent advantages for early reasoning tokens, paradoxically underperforming outcome-only GRPO. We propose Skip-Connected Optimization (SKPO), which decomposes reasoning into upstream and downstream phases: upstream receives dense rewards from downstream Monte Carlo sampling with single-stream optimization; downstream maintains group-relative optimization, where a skip connection concatenates the upstream segment with the original problem, enabling the model to leverage helpful upstream reasoning while preserving the freedom to bypass flawed reasoning through direct problem access. Experiments demonstrate improvements of 3.91% and 6.17% relative gains over the strongest baselines on Qwen2.5-Math-7B and Llama-3.2-3B respectively across mathematical benchmarks and out-of-domain tasks including general reasoning and code generation. Further analysis reveals an implicit advantage: SKPO generates trajectories with higher intermediate-step quality even when matched for final correctness. △ Less

Submitted 9 April, 2026; originally announced April 2026.

PRISM: Dynamic Primitive-Based Forecasting for Large-Scale GPU Cluster Workloads

Authors: Xin Wu, Fei Teng, Xingwang Li, Bin Zheng, Qiang Duan

Abstract: Accurately forecasting GPU workloads is essential for AI infrastructure, enabling efficient scheduling, resource allocation, and power management. Modern workloads are highly volatile, multiple periodicity, and heterogeneous, making them challenging for traditional predictors. We propose PRISM, a primitive-based compositional forecasting framework combining dictionary-driven temporal decomposition… ▽ More Accurately forecasting GPU workloads is essential for AI infrastructure, enabling efficient scheduling, resource allocation, and power management. Modern workloads are highly volatile, multiple periodicity, and heterogeneous, making them challenging for traditional predictors. We propose PRISM, a primitive-based compositional forecasting framework combining dictionary-driven temporal decomposition with adaptive spectral refinement. This dual representation extracts stable, interpretable workload signatures across diverse GPU jobs. Evaluated on large-scale production traces, PRISM achieves state-of-the-art results. It significantly reduces burst-phase errors, providing a robust, architecture-aware foundation for dynamic resource management in GPU-powered AI platforms. △ Less

Submitted 26 March, 2026; originally announced March 2026.

Comments: Accepted by DAC'26

Journal ref: 979-8-4007-2254-7/2026/07

Flow-based Polynomial Chaos Expansion for Uncertainty Quantification in Power System Dynamic Simulation

Authors: Le Fang, Wangkun Xu, Fei Teng

Abstract: The large-scale integration of renewable energy sources introduces significant operational uncertainty into power systems. Although Polynomial Chaos Expansion (PCE) provides an efficient tool for uncertainty quantification (UQ) in power system dynamics, its accuracy depends critically on the faithful representation of input uncertainty, an assumption that is oftern violated in practice due to corr… ▽ More The large-scale integration of renewable energy sources introduces significant operational uncertainty into power systems. Although Polynomial Chaos Expansion (PCE) provides an efficient tool for uncertainty quantification (UQ) in power system dynamics, its accuracy depends critically on the faithful representation of input uncertainty, an assumption that is oftern violated in practice due to correlated, non-Gaussian, and otherwise complex data distributions. In contrast to purely data-driven surrogates that often overlook rigorous input distribution modelling, this paper introduces flow-based PCE, a unified framework that couples expressive input modelling with efficient uncertainty propagation. Specifically, normalising flows are employed to learn an invertible transport map from a simple base distribution to the empirical joint distribution of uncertain inputs, and this map is then integrated directly into the PCE construction. In addition, the Map Smoothness Index (MSI) is introduced as a new metric to quantify the quality of the learned map, and smoother transformations are shown to yield more accurate PCE surrogates. The proposed Flow-based PCE framework is validated on benchmark dynamic models, including the IEEE 14-bus system and the Great Britain transmission system, under a range of uncertainty scenarios. △ Less

Submitted 20 March, 2026; originally announced March 2026.

Stochastic Virtual Power Plant Dispatch via Temporally Aggregated Distributed Predictive Control with Performance Guarantees

Authors: Luca Santosuosso, Fei Teng, Sonja Wogrin

Abstract: This paper addresses the energy dispatch of a virtual power plant comprising renewable generation, energy storage, and thermal units under uncertainty in renewable output, energy prices, and energy demand. The nonlinear dynamics and multiple sources of uncertainty render traditional stochastic model predictive control (MPC) computationally intractable as the dispatch horizon, scenario set, and ass… ▽ More This paper addresses the energy dispatch of a virtual power plant comprising renewable generation, energy storage, and thermal units under uncertainty in renewable output, energy prices, and energy demand. The nonlinear dynamics and multiple sources of uncertainty render traditional stochastic model predictive control (MPC) computationally intractable as the dispatch horizon, scenario set, and asset portfolio expand. To overcome this limitation, we propose a novel controller that seamlessly integrates MPC with time series aggregation and distributed optimization, simultaneously reducing the temporal, asset, and scenario dimensions of the problem. The resulting controller provides a rigorous performance guarantee through theoretically validated bounds on its approximation error, while leveraging dual information from previous MPC iterations to adaptively optimize the temporal aggregation. Numerical results show that the proposed controller reduces runtime by over 50% relative to traditional stochastic MPC and, crucially, restores tractability where the full-scale dispatch model proves intractable. △ Less

Submitted 19 March, 2026; originally announced March 2026.

Panoramic Multimodal Semantic Occupancy Prediction for Quadruped Robots

Authors: Guoqiang Zhao, Zhe Yang, Sheng Wu, Fei Teng, Mengfei Duan, Yuanfan Zheng, Kai Luo, Kailun Yang

Abstract: Panoramic imagery provides holistic 360° visual coverage for environmental perception in quadruped robots. However, existing occupancy prediction methods are primarily designed for wheeled autonomous driving and rely heavily on RGB cues, which limits their robustness in complex, dynamically changing environments. To bridge this gap, we introduce PanoMMOcc, the first real-world panoramic multimodal… ▽ More Panoramic imagery provides holistic 360° visual coverage for environmental perception in quadruped robots. However, existing occupancy prediction methods are primarily designed for wheeled autonomous driving and rely heavily on RGB cues, which limits their robustness in complex, dynamically changing environments. To bridge this gap, we introduce PanoMMOcc, the first real-world panoramic multimodal occupancy dataset for quadruped robots, comprising four sensing modalities collected across diverse scenes. We further propose VoxelHound, a panoramic multimodal occupancy perception framework tailored to legged locomotion and spherical imaging. VoxelHound incorporates a Vertical Jitter Compensation (VJC) module to mitigate severe viewpoint perturbations caused by body pitch and roll during locomotion, enabling more consistent spatial reasoning, and a Multimodal Information Prompt Fusion (MIPF) module to effectively integrate panoramic visual cues with auxiliary modalities for enhanced volumetric occupancy prediction. We also establish a comprehensive benchmark on PanoMMOcc and provide detailed dataset analyses to enable systematic evaluation in challenging embodied perception scenarios. Extensive experiments demonstrate that VoxelHound achieves state-of-the-art performance on PanoMMOcc, with a +4.16 gain in mIoU. The dataset and code will be publicly released to facilitate future research on panoramic multimodal 3D perception for embodied robotic systems at https://github.com/SXDR/PanoMMOcc. △ Less

Submitted 7 August, 2026; v1 submitted 13 March, 2026; originally announced March 2026.

Comments: The dataset and code will be publicly released at https://github.com/SXDR/PanoMMOcc

O3N: Omnidirectional Open-Vocabulary Occupancy Prediction for Urban Autonomous Agents

Authors: Mengfei Duan, Hao Shi, Fei Teng, Guoqiang Zhao, Yuheng Zhang, Zhiyong Li, Kailun Yang

Abstract: Understanding and reconstructing the 3D world through omnidirectional perception is becoming increasingly important for intelligent transportation and urban autonomous agents operating in shared spaces. However, existing vision-based 3D occupancy prediction methods are constrained by limited perspective inputs and a predefined training distribution, making them difficult to apply to embodied agent… ▽ More Understanding and reconstructing the 3D world through omnidirectional perception is becoming increasingly important for intelligent transportation and urban autonomous agents operating in shared spaces. However, existing vision-based 3D occupancy prediction methods are constrained by limited perspective inputs and a predefined training distribution, making them difficult to apply to embodied agents that require comprehensive and safe perception of scenes in open-world exploration. To address this, we present O3N, the first framework for open-vocabulary occupancy prediction from a single omnidirectional RGB image. O3N embeds omnidirectional voxels in a polar-spiral topology via the Polar-spiral Mamba (PsM) module, enabling continuous spatial representation and long-range context modeling across 360°. The Occupancy Cost Aggregation (OCA) module introduces a principled mechanism for unifying geometric and semantic supervision within the voxel space, ensuring consistency between the reconstructed geometry and the underlying semantic structure. Moreover, Natural Modality Alignment (NMA) establishes a gradient-free alignment pathway that harmonizes visual features, voxel embeddings, and text semantics, forming a consistent ``pixel-voxel-text'' representation triad. Extensive experiments on multiple models demonstrate that our method not only achieves state-of-the-art performance on QuadOcc and Human360Occ benchmarks but also exhibits remarkable cross-scene generalization and semantic scalability, highlighting the potential of O3N for scalable open-world 3D scene understanding. The source code will be made publicly available at https://github.com/MengfeiD/O3N △ Less

Submitted 23 July, 2026; v1 submitted 12 March, 2026; originally announced March 2026.

Comments: The source code will be made publicly available at https://github.com/MengfeiD/O3N

Distribution Shift Is Key to Learning Invariant Prediction

Authors: Hong Zheng, Fei Teng

Abstract: An interesting phenomenon arises: Empirical Risk Minimization (ERM) sometimes outperforms methods specifically designed for out-of-distribution tasks. This motivates an investigation into the reasons behind such behavior beyond algorithmic design. In this study, we find that one such reason lies in the distribution shift across training domains. A large degree of distribution shift can lead to bet… ▽ More An interesting phenomenon arises: Empirical Risk Minimization (ERM) sometimes outperforms methods specifically designed for out-of-distribution tasks. This motivates an investigation into the reasons behind such behavior beyond algorithmic design. In this study, we find that one such reason lies in the distribution shift across training domains. A large degree of distribution shift can lead to better performance even under ERM. Specifically, we derive several theoretical and empirical findings demonstrating that distribution shift plays a crucial role in model learning and benefits learning invariant prediction. Firstly, the proposed upper bounds indicate that the degree of distribution shift directly affects the prediction ability of the learned models. If it is large, the models' ability can increase, approximating invariant prediction models that make stable predictions under arbitrary known or unseen domains; and vice versa. We also prove that, under certain data conditions, ERM solutions can achieve performance comparable to that of invariant prediction models. Secondly, the empirical validation results demonstrated that the predictions of learned models approximate those of Oracle or Optimal models, provided that the degree of distribution shift in the training data increases. △ Less

Submitted 18 January, 2026; originally announced January 2026.

Headroom as A Grid Service in Software-Defined Power Grids: A Peak-to-Peak Control Design Approach

Authors: Zhongda Chu, Fei Teng

Abstract: To address system frequency challenges driven by the integration of renewable generation, advanced control strategies are designed at the device level to provide effective frequency support following disturbances. However, typically relying on energy-based performance metrics, these methods cannot guarantee the system frequency constraints such as frequency nadir and maximum Rate-of-Change-of-Freq… ▽ More To address system frequency challenges driven by the integration of renewable generation, advanced control strategies are designed at the device level to provide effective frequency support following disturbances. However, typically relying on energy-based performance metrics, these methods cannot guarantee the system frequency constraints such as frequency nadir and maximum Rate-of-Change-of-Frequency (RoCoF). Moreover, locally-designed frequency support cannot minimize the overall system cost to maintain frequency stability. On the other hand, the concept of frequency-constrained system scheduling is introduced, which incorporates frequency dynamic constraints into the system economic optimization, so that frequency requirements can be maintained with minimum cost. However, these works rely on analytical approximations of the frequency dynamic metrics, which are mathematically complicated and tend to be over-conservative for the approximation of IBR headroom requirements. △ Less

Submitted 14 December, 2025; originally announced December 2025.

Do Not Merge My Model! Safeguarding Open-Source LLMs Against Unauthorized Model Merging

Authors: Qinfeng Li, Miao Pan, Jintao Chen, Fu Teng, Zhiqiang Shen, Ge Su, Hao Peng, Xuhong Zhang

Abstract: Model merging has emerged as an efficient technique for expanding large language models (LLMs) by integrating specialized expert models. However, it also introduces a new threat: model merging stealing, where free-riders exploit models through unauthorized model merging. Unfortunately, existing defense mechanisms fail to provide effective protection. Specifically, we identify three critical protec… ▽ More Model merging has emerged as an efficient technique for expanding large language models (LLMs) by integrating specialized expert models. However, it also introduces a new threat: model merging stealing, where free-riders exploit models through unauthorized model merging. Unfortunately, existing defense mechanisms fail to provide effective protection. Specifically, we identify three critical protection properties that existing methods fail to simultaneously satisfy: (1) proactively preventing unauthorized merging; (2) ensuring compatibility with general open-source settings; (3) achieving high security with negligible performance loss. To address the above issues, we propose MergeBarrier, a plug-and-play defense that proactively prevents unauthorized merging. The core design of MergeBarrier is to disrupt the Linear Mode Connectivity (LMC) between the protected model and its homologous counterparts, thereby eliminating the low-loss path required for effective model merging. Extensive experiments show that MergeBarrier effectively prevents model merging stealing with negligible accuracy loss. △ Less

Submitted 20 November, 2025; v1 submitted 13 November, 2025; originally announced November 2025.

Comments: Accepted by AAAI 2026 Conference

Asymptotic Simplicity and Scattering in General Relativity from Quantum Field Theory

Authors: Stefano De Angelis, Aidan Herderschee, Radu Roiban, Fei Teng

Abstract: We investigate the fate of asymptotic simplicity in physically relevant settings of compact-object scattering. Using the stress tensor of a two-body system as a source, we compute the spacetime metric in General Relativity at finite observer distance in an asymptotic expansion. To do so, we relate the metric to the final-state graviton one-point function in momentum space, which is computed using… ▽ More We investigate the fate of asymptotic simplicity in physically relevant settings of compact-object scattering. Using the stress tensor of a two-body system as a source, we compute the spacetime metric in General Relativity at finite observer distance in an asymptotic expansion. To do so, we relate the metric to the final-state graviton one-point function in momentum space, which is computed using perturbative QFT techniques. Both the simple pole and the infrared-related logarithmic branch cut in the virtuality of the external graviton contribute nontrivially. We focus on determining the fall-off behavior of the Newman-Penrose scalars, confirming previous predictions that Sachs's peeling property is violated at leading order in the post-Minkowski expansion. Our analysis at higher orders in the post-Minkowskian expansion reveals a significantly stronger breakdown of the peeling property than previously recognized, which is the result of nonlinear, long-range interactions between localized sources and the surrounding gravitational field. △ Less

Submitted 12 March, 2026; v1 submitted 13 November, 2025; originally announced November 2025.

Comments: 41 pages + references; v2: references added, typos corrected, improved discussion

AGRAG: Advanced Graph-based Retrieval-Augmented Generation for LLMs

Authors: Yubo Wang, Haoyang Li, Fei Teng, Lei Chen

Abstract: Graph-based retrieval-augmented generation (Graph-based RAG) has demonstrated significant potential in enhancing Large Language Models (LLMs) with structured knowledge. However, existing methods face three critical challenges: Inaccurate Graph Construction, caused by LLM hallucination; Poor Reasoning Ability, caused by failing to generate explicit reasons telling LLM why certain chunks were select… ▽ More Graph-based retrieval-augmented generation (Graph-based RAG) has demonstrated significant potential in enhancing Large Language Models (LLMs) with structured knowledge. However, existing methods face three critical challenges: Inaccurate Graph Construction, caused by LLM hallucination; Poor Reasoning Ability, caused by failing to generate explicit reasons telling LLM why certain chunks were selected; and Inadequate Answering, which only partially answers the query due to the inadequate LLM reasoning, making their performance lag behind NaiveRAG on certain tasks. To address these issues, we propose AGRAG, an advanced graph-based retrieval-augmented generation framework. When constructing the graph, AGRAG substitutes the widely used LLM entity extraction method with a statistics-based method, avoiding hallucination and error propagation. During retrieval, AGRAG formulates the graph reasoning procedure as the Minimum Cost Maximum Influence (MCMI) subgraph generation problem, where we try to include more nodes with high influence score, but with less involving edge cost, to make the generated reasoning paths more comprehensive. We prove this problem to be NP-hard, and propose a greedy algorithm to solve it. The MCMI subgraph generated can serve as explicit reasoning paths to tell LLM why certain chunks were retrieved, thereby making the LLM better focus on the query-related part contents of the chunks, reducing the impact of noise, and improving AGRAG's reasoning ability. Furthermore, compared with the simple tree-structured reasoning paths, our MCMI subgraph can allow more complex graph structures, such as cycles, and improve the comprehensiveness of the generated reasoning paths. The code and prompt of AGRAG are released at: https://github.com/Wyb0627/AGRAG. △ Less

Submitted 17 March, 2026; v1 submitted 2 November, 2025; originally announced November 2025.

Comments: ICDE 2026 Camera-ready

Adaptive Federated Learning to Optimize Integrated Flows in Cyber-Physical Data Centers

Authors: Junhong Liu, Lanxin Du, Yujia Li, Rong-Peng Liu, Yunfeng Li, Fei Teng, Francis Yunhe Hou

Abstract: Data centers play an increasingly critical role in societal digitalization, yet their rapidly growing energy demand poses significant challenges for sustainable operation. To enhance the energy efficiency of geographically distributed data centers, this paper formulates a multi-period optimization model that captures the interdependence of electricity, heat, and data flows. The optimization of suc… ▽ More Data centers play an increasingly critical role in societal digitalization, yet their rapidly growing energy demand poses significant challenges for sustainable operation. To enhance the energy efficiency of geographically distributed data centers, this paper formulates a multi-period optimization model that captures the interdependence of electricity, heat, and data flows. The optimization of such integrated multi-domain flows inherently involves mixed-integer formulations and the access to proprietary or sensitive datasets, which correspondingly exacerbate computational complexity and raise data-privacy concerns. To address these challenges, an adaptive federated learning-to-optimization approach is proposed, accounting for the heterogeneity of datasets across distributed data centers. To safeguard privacy, cryptography techniques are leveraged in both the learning and optimization processes. A model acceptance criterion with convergence guarantee is developed to improve learning performance and filter out potentially contaminated data, while a verifiable double aggregation mechanism is further proposed to simultaneously ensure privacy and integrity of shared data during optimization. Theoretical analysis and numerical simulations demonstrate that the proposed approach preserves the privacy and integrity of shared data, achieves near-optimal performance, and exhibits high computational efficiency, making it suitable for large-scale data center optimization under privacy constraints. △ Less

Submitted 14 July, 2026; v1 submitted 1 November, 2025; originally announced November 2025.

OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback

Authors: Kai Luo, Hao Shi, Kunyu Peng, Fei Teng, Sheng Wu, Kaiwei Wang, Kailun Yang

Abstract: To address panoramic distortion, large search space, and identity ambiguity under a 360° FoV, OmniTrack++ adopts a feedback-driven framework that progressively refines perception with trajectory cues. A DynamicSSM block first stabilizes panoramic features, implicitly alleviating geometric distortion. On top of normalized representations, FlexiTrack Instances use trajectory-informed feedback for fl… ▽ More To address panoramic distortion, large search space, and identity ambiguity under a 360° FoV, OmniTrack++ adopts a feedback-driven framework that progressively refines perception with trajectory cues. A DynamicSSM block first stabilizes panoramic features, implicitly alleviating geometric distortion. On top of normalized representations, FlexiTrack Instances use trajectory-informed feedback for flexible localization and reliable short-term association. To ensure long-term robustness, an ExpertTrack Memory consolidates appearance cues via a Mixture-of-Experts design, enabling recovery from fragmented tracks and reducing identity drift. Finally, a Tracklet Management module adaptively switches between end-to-end and tracking-by-detection modes according to scene dynamics, offering a balanced and scalable solution for panoramic MOT. To support rigorous evaluation, we establish the EmboTrack benchmark, a comprehensive dataset for panoramic MOT that includes QuadTrack, captured with a quadruped robot, and BipTrack, collected with a bipedal wheel-legged robot. Together, these datasets span wide-angle environments and diverse motion patterns, providing a challenging testbed for real-world panoramic perception. Extensive experiments on JRDB and EmboTrack demonstrate that OmniTrack++ achieves state-of-the-art performance, yielding substantial HOTA improvements of +3.94 on JRDB and +15.03 on QuadTrack over the original OmniTrack. These results highlight the effectiveness of trajectory-informed feedback, adaptive paradigm switching, and robust long-term memory in advancing panoramic multi-object tracking. Datasets and code will be made available at https://github.com/xifen523/OmniTrack. △ Less

Submitted 4 May, 2026; v1 submitted 1 November, 2025; originally announced November 2025.

Comments: Extended version of CVPR 2025 paper arXiv:2503.04565. Datasets and code will be made publicly available at https://github.com/xifen523/OmniTrack

InteractComp: Evaluating Search Agents With Ambiguous Queries

Authors: Mingyi Deng, Lijun Huang, Yani Fan, Fanqi Kong, Jiayi Zhang, Fashen Ren, Jinyi Bai, Fuzhen Yang, Dayi Miao, Zhaoyang Yu, Yifan Wu, Yanfei Zhang, Fengwei Teng, Yingjia Wan, Song Hu, Yude Li, Xin Jin, Conghao Hu, Haoyu Li, Qirui Fu, Tai Zhong, Xinyu Wang, Xiangru Tang, Nan Tang, Chenglin Wu , et al. (1 additional authors not shown)

Abstract: Language agents have demonstrated remarkable potential in web search and information retrieval. However, many search-agent benchmarks assume that user queries are complete and unambiguous. This assumption leaves under-tested a practical failure mode: agents may face ambiguous requests where the intended target cannot be identified without clarification. Yet most agents lack interactive mechanisms… ▽ More Language agents have demonstrated remarkable potential in web search and information retrieval. However, many search-agent benchmarks assume that user queries are complete and unambiguous. This assumption leaves under-tested a practical failure mode: agents may face ambiguous requests where the intended target cannot be identified without clarification. Yet most agents lack interactive mechanisms during the search process, and existing benchmarks cannot assess this capability. To address this gap, we introduce InteractComp, a benchmark designed to evaluate whether search agents can recognize query ambiguity and actively interact to resolve it during search. Following the principle of easy to verify, interact to disambiguate, we construct 210 expert-curated questions across 9 domains through a target-distractor methodology that creates controlled ambiguity resolvable only through interaction. Evaluation of 17 models reveals striking failure: the best model achieves only 13.73% accuracy despite 71.50% with complete context, exposing systematic overconfidence rather than reasoning deficits. Forced interaction produces dramatic gains, demonstrating latent capability current strategies fail to engage. Longitudinal analysis shows interaction capabilities stagnated over 15 months while search performance improved seven-fold, revealing a critical blind spot. This stagnation, coupled with the immediate feedback inherent to search tasks, makes InteractComp a valuable resource for both evaluating and training interaction capabilities in search agents. The code is available at https://github.com/FoundationAgents/InteractComp. △ Less

Submitted 24 July, 2026; v1 submitted 28 October, 2025; originally announced October 2025.

Comments: Accepted at the 43rd International Conference on Machine Learning (ICML 2026), Seoul, South Korea. 9 pages, 4 figures

Data-driven Communication and Control Design for Distributed Frequency Regulation with Black-box Inverters

Authors: Michael Nestor, Jiaxin Wang, Ning Zhang, Fei Teng

Abstract: The increasing penetration of inverter-based resources into the power grid, with often only black-box models available, challenges long-standing frequency control methods. Most recent works take a decentralized approach without online device coordination via communication. This paper considers both dynamic behavior and communication within secondary frequency control on an intermediate timescale.… ▽ More The increasing penetration of inverter-based resources into the power grid, with often only black-box models available, challenges long-standing frequency control methods. Most recent works take a decentralized approach without online device coordination via communication. This paper considers both dynamic behavior and communication within secondary frequency control on an intermediate timescale. We develop a distributed data-driven approach that utilizes peer-to-peer communication between inverters to avoid the need for a central control center. To enable a trade off between communication network requirements and control performance, we present a framework to guide communication topology design for secondary frequency regulation. Following design of the inter-agent information exchange scheme, we design a controller that is structured according to the communication topology with a closed-loop stability guarantee. Case studies on the IEEE 39-bus system validate the framework and illustrate the trade-off between communication requirements and control performance that is enabled by our approach. △ Less

Submitted 20 October, 2025; originally announced October 2025.

Comments: Preprint submitted to PSCC 2026

Eddy viscosity by Lévy transport noises

Authors: Dejun Luo, Feifan Teng

Abstract: We consider stochastic 2D Euler equations with $L^2$-initial vorticity and driven by Lévy transport noise in the Marcus sense. Under a suitable scaling limit of the noises, we prove that the weak solutions converge weakly to the unique solution of the deterministic 2D Navier-Stokes equation. This shows that small scale jump noises generate eddy viscosity, extending the recent studies on Itô-Strato… ▽ More We consider stochastic 2D Euler equations with $L^2$-initial vorticity and driven by Lévy transport noise in the Marcus sense. Under a suitable scaling limit of the noises, we prove that the weak solutions converge weakly to the unique solution of the deterministic 2D Navier-Stokes equation. This shows that small scale jump noises generate eddy viscosity, extending the recent studies on Itô-Stratonovich diffusion limit to discontinuous setting. △ Less

Submitted 15 October, 2025; originally announced October 2025.

Comments: 41 pages

The Debate on RLVR Reasoning Capability Boundary: Shrinkage, Expansion, or Both? A Two-Stage Dynamic View

Authors: Xinhao Yao, Lu Yu, Xiaolin Hu, Fengwei Teng, Qing Cui, Jun Zhou, Yong Liu

Abstract: The ongoing debate on whether reinforcement learning with verifiable rewards (RLVR) expands or shrinks the reasoning capabilities of large language models (LLMs) remains unresolved. Some studies contend that RLVR mainly improves sampling efficiency but at the expense of diversity and exploratory capacity, resulting in capability boundary shrinkage. In contrast, others demonstrate that prolonged tr… ▽ More The ongoing debate on whether reinforcement learning with verifiable rewards (RLVR) expands or shrinks the reasoning capabilities of large language models (LLMs) remains unresolved. Some studies contend that RLVR mainly improves sampling efficiency but at the expense of diversity and exploratory capacity, resulting in capability boundary shrinkage. In contrast, others demonstrate that prolonged training can lead to the emergence of novel reasoning strategies, suggesting capability boundary expansion. To reconcile these contradictory findings, we theoretically and empirically show that both perspectives are partially valid-each aligning with a separate phase in an inherent two-stage probability mass dynamic: (1) Exploitation stage: initially, the model primarily samples explored high-reward and low-reward tokens, while rarely selecting the potentially optimal token. Positive advantage estimates increase the probability of high-reward tokens and decrease those of low-reward tokens, yet the optimal token's probability remains largely unchanged during this stage. (2) Exploration stage: as training advances, the growth rate of previously acquired high-reward tokens slows as their probabilities approach saturation. When a potentially optimal token-now receiving positive advantage estimates-is occasionally sampled, its probability increases, while those of the originally high-reward tokens decrease. This dynamic suggests that over-exploitation during the exploitation stage may lead to capability boundary shrinkage, whereas prolonged training into the exploration stage can promote an expansion of the reasoning capability boundary. Building upon our insights, we revisit the potential of only using relative negative gradients for prolonging training, providing a theoretical and empirical foundation for the development of more advanced reasoning capabilities. △ Less

Submitted 5 October, 2025; originally announced October 2025.

DepTR-MOT: Unveiling the Potential of Depth-Informed Trajectory Refinement for Multi-Object Tracking

Authors: Buyin Deng, Lingxin Huang, Kai Luo, Fei Teng, Kailun Yang

Abstract: Visual Multi-Object Tracking (MOT) is a crucial component of robotic perception, yet existing Tracking-By-Detection (TBD) methods often rely on 2D cues, such as bounding boxes and motion modeling, which struggle under occlusions and close-proximity interactions. Trackers relying on these 2D cues are particularly unreliable in robotic environments, where dense targets and frequent occlusions are co… ▽ More Visual Multi-Object Tracking (MOT) is a crucial component of robotic perception, yet existing Tracking-By-Detection (TBD) methods often rely on 2D cues, such as bounding boxes and motion modeling, which struggle under occlusions and close-proximity interactions. Trackers relying on these 2D cues are particularly unreliable in robotic environments, where dense targets and frequent occlusions are common. While depth information has the potential to alleviate these issues, most existing MOT datasets lack depth annotations, leading to its underexploited role in the domain. To unveil the potential of depth-informed trajectory refinement, we introduce DepTR-MOT, a DETR-based detector enhanced with instance-level depth information. Specifically, we propose two key innovations: (i) foundation model-based instance-level soft depth label supervision, which refines depth prediction, and (ii) the distillation of dense depth maps to maintain global depth consistency. These strategies enable DepTR-MOT to output instance-level depth during inference, without requiring foundation models and without additional computational cost. By incorporating depth cues, our method enhances the robustness of the TBD paradigm, effectively resolving occlusion and close-proximity challenges. Experiments on both the QuadTrack and DanceTrack datasets demonstrate the effectiveness of our approach, achieving HOTA scores of 27.59 and 44.47, respectively. In particular, results on QuadTrack, a robotic platform MOT dataset, highlight the advantages of our method in handling occlusion and close-proximity challenges in robotic tracking. The source code will be made publicly available at https://github.com/warriordby/DepTR-MOT. △ Less

Submitted 21 September, 2025; originally announced September 2025.

Comments: The source code will be made publicly available at https://github.com/warriordby/DepTR-MOT

Visually Grounded Narratives: Reducing Cognitive Burden in Researcher-Participant Interaction

Authors: Runtong Wu, Jiayao Song, Fei Teng, Xianhao Ren, Yuyan Gao, Kailun Yang

Abstract: Narrative inquiry has been one of the prominent application domains for the analysis of human experience, aiming to know more about the complexity of human society. However, researchers are often required to transform various forms of data into coherent hand-drafted narratives in storied form throughout narrative analysis, which brings an immense burden of data analysis. Participants, too, are exp… ▽ More Narrative inquiry has been one of the prominent application domains for the analysis of human experience, aiming to know more about the complexity of human society. However, researchers are often required to transform various forms of data into coherent hand-drafted narratives in storied form throughout narrative analysis, which brings an immense burden of data analysis. Participants, too, are expected to engage in member checking and presentation of these narrative products, which involves reviewing and responding to large volumes of documents. Given the dual burden and the need for more efficient and participant-friendly approaches to narrative making and representation, we made a first attempt: (i) a new paradigm is proposed, NAME, as the initial attempt to push the field of narrative inquiry. Name is able to transfer research documents into coherent story images, alleviating the cognitive burden of interpreting extensive text-based materials during member checking for both researchers and participants. (ii) We develop an actor location and shape module to facilitate plausible image generation. (iii) We have designed a set of robust evaluation metrics comprising three key dimensions to objectively measure the perceptual quality and narrative consistency of generated characters. Our approach consistently demonstrates state-of-the-art performance across different data partitioning schemes. Remarkably, while the baseline relies on the full 100% of the available data, our method requires only 0.96% yet still reduces the FID score from 195 to 152. Under identical data volumes, our method delivers substantial improvements: for the 70:30 split, the FID score decreases from 175 to 152, and for the 95:5 split, it is nearly halved from 96 to 49. Furthermore, the proposed model achieves a score of 3.62 on the newly introduced metric, surpassing the baseline score of 2.66. △ Less

Submitted 30 August, 2025; originally announced September 2025.

Privacy, Informed Consent and the Demand for Anonymisation of Smart Meter Data

Authors: Saurab Chhachhi, Fei Teng

Abstract: Access to smart meter data offers system-wide benefits but raises significant privacy concerns due to the personal information it contains. Privacy-preserving techniques could facilitate wider access, though they introduce privacy-utility trade-offs. Understanding consumer valuations for anonymisation can help identify appropriate trade-offs. However, existing studies do not focus on anonymisation… ▽ More Access to smart meter data offers system-wide benefits but raises significant privacy concerns due to the personal information it contains. Privacy-preserving techniques could facilitate wider access, though they introduce privacy-utility trade-offs. Understanding consumer valuations for anonymisation can help identify appropriate trade-offs. However, existing studies do not focus on anonymisation specifically or account for information asymmetries regarding privacy risks, raising questions about the validity of informed consent under current regulations. We use a mixed-methods approach to estimate non-monetary (willingness-to-share and smart metering demand) and monetary (willingness-to-pay/accept) preferences for anonymisation, based on a representative sample of 965 GB bill payers. An embedded randomised control trial examines the effect of providing information about privacy implications. On average, consumers are willing to pay for anonymisation, are more willing to share data when anonymised and less willing to share non-anonymised data once anonymisation is presented as an option. However, a significant minority remains unwilling to adopt smart meters, despite anonymisation. We find strong evidence of information asymmetries that suppress demand for anonymisation and identify substantial variation across demographic and electricity supply characteristics. Qualitative responses corroborate the quantitative findings, underscoring the need for stronger privacy defaults, user-centric design, and consent mechanisms that enable truly informed decisions. △ Less

Submitted 27 August, 2025; originally announced September 2025.

Comments: 24 pages, 7 figures, 44 page appendix

VERIRL: Boosting the LLM-based Verilog Code Generation via Reinforcement Learning

Authors: Fu Teng, Miao Pan, Xuhong Zhang, Zhezhi He, Yiyao Yang, Xinyi Chai, Mengnan Qi, Liqiang Lu, Jianwei Yin

Abstract: Recent advancements in code generation have shown remarkable success across software domains, yet hardware description languages (HDLs) such as Verilog remain underexplored due to their concurrency semantics, syntactic rigidity, and simulation complexity. In this work, we address these challenges by introducing a reinforcement learning (RL) framework tailored for Verilog code generation. We first… ▽ More Recent advancements in code generation have shown remarkable success across software domains, yet hardware description languages (HDLs) such as Verilog remain underexplored due to their concurrency semantics, syntactic rigidity, and simulation complexity. In this work, we address these challenges by introducing a reinforcement learning (RL) framework tailored for Verilog code generation. We first construct Veribench-53K, a high-quality dataset curated from over 700K Verilog problems, enriched with structured prompts, complexity labels, and diverse testbenches. To tackle the problem of sparse and noisy reward signals, we propose a Trace-back based Rescore mechanism that leverages reasoning paths and iterative refinement to enhance feedback reliability and support reward model training. Furthermore, to mitigate catastrophic forgetting and overfitting during RL fine-tuning, we introduce a sample-balanced weighting strategy that adaptively balances learning dynamics based on reward-probability distributions. These innovations are integrated into an iterative RL pipeline that co-evolves the policy and reward models. In contrast to recent work such as CraftRTL, which relies on large-scale closed-source model distillation, and DeepSeek-style approaches that struggle with sparse feedback, our method demonstrates superior performance using a smaller but high-quality dataset combined with RL optimization. Experiments on Verilog generation tasks demonstrate state-of-the-art performance, with substantial gains in test pass rate, functional correctness, and compilation robustness. Our findings highlight the potential of RL-driven approaches for structured code generation in hardware-centric domains. VERIRL is publicly available at https://github.com/omniAI-Lab/VeriRL. △ Less

Submitted 25 August, 2025; originally announced August 2025.

ERIS: An Energy-Guided Feature Disentanglement Framework for Out-of-Distribution Time Series Classification

Authors: Xin Wu, Fei Teng, Ji Zhang, Xingwang Li, Yuxuan Liang

Abstract: An ideal time series classification (TSC) should be able to capture invariant representations, but achieving reliable performance on out-of-distribution (OOD) data remains a core obstacle. This obstacle arises from the way models inherently entangle domain-specific and label-relevant features, resulting in spurious correlations. While feature disentanglement aims to solve this, current methods are… ▽ More An ideal time series classification (TSC) should be able to capture invariant representations, but achieving reliable performance on out-of-distribution (OOD) data remains a core obstacle. This obstacle arises from the way models inherently entangle domain-specific and label-relevant features, resulting in spurious correlations. While feature disentanglement aims to solve this, current methods are largely unguided, lacking the semantic direction required to isolate truly universal features. To address this, we propose an end-to-end Energy-Regularized Information for Shift-Robustness (ERIS) framework to enable guided and reliable feature disentanglement. The core idea is that effective disentanglement requires not only mathematical constraints but also semantic guidance to anchor the separation process. ERIS incorporates three key mechanisms to achieve this goal. Specifically, we first introduce an energy-guided calibration mechanism, which provides crucial semantic guidance for the separation, enabling the model to self-calibrate. Additionally, a weight-level orthogonality strategy enforces structural independence between domain-specific and label-relevant features, thereby mitigating their interference. Moreover, an auxiliary adversarial generalization mechanism enhances robustness by injecting structured perturbations. Experiments across four benchmarks demonstrate that ERIS achieves a statistically significant improvement over state-of-the-art baselines, consistently securing the top performance rank. △ Less

Submitted 26 September, 2025; v1 submitted 19 August, 2025; originally announced August 2025.

Journal ref: Information Fusion 135, 104407 (2026)

LLMLog: Advanced Log Template Generation via LLM-driven Multi-Round Annotation

Authors: Fei Teng, Haoyang Li, Lei Chen

Abstract: Modern computing systems, such as HDFS and Spark, produce vast quantities of logs that developers use for tasks like anomaly detection and error analysis. To simplify log analysis, template generation methods have been proposed to standardize log formats, transforming unstructured data into structured templates. Existing heuristic-based methods and neural network-based methods suffer from low accu… ▽ More Modern computing systems, such as HDFS and Spark, produce vast quantities of logs that developers use for tasks like anomaly detection and error analysis. To simplify log analysis, template generation methods have been proposed to standardize log formats, transforming unstructured data into structured templates. Existing heuristic-based methods and neural network-based methods suffer from low accuracy problems due to the reliance on handcrafted heuristics or specific log patterns in training sets. Recently, large language models (LLMs) have shown great potential in log template generation. However, they often struggle with ambiguous, complex, or highly specific log content, which can lead to errors in generating accurate templates. To address these challenges, we propose LLMLog, a multi-round annotation framework with adaptive in-context learning. We first propose an edit-distance-based similarity metric to evaluate log similarity. Then, we introduce a method to select the most informative $k$ unlabeled logs for annotation by considering both the representativeness of the logs and the confidence of LLM predictions. Additionally, we design an adaptive context selection strategy that adaptively selects labeled logs to ensure comprehensive keyword coverage for unlabeled logs. These labeled logs serve as the context for LLMs to better understand the unlabeled logs, thereby enhancing the accuracy of template generation. Extensive experiments on sixteen datasets demonstrate that LLMLog outperforms the state-of-the-art approaches. △ Less

Submitted 13 August, 2025; originally announced August 2025.

Comments: Accepted in VLDB 2025

Learning-Enabled Adaptive Power Capping Scheme for Cloud Data Centers

Authors: Yimeng Sun, Zhaohao Ding, Payman Dehghanian, Fei Teng

Abstract: The rapid growth of the digital economy and artificial intelligence has transformed cloud data centers into essential infrastructure with substantial energy consumption and carbon emission, necessitating effective energy management. However, existing methods face challenges such as incomplete information, uncertain parameters, and dynamic environments, which hinder their real-world implementation.… ▽ More The rapid growth of the digital economy and artificial intelligence has transformed cloud data centers into essential infrastructure with substantial energy consumption and carbon emission, necessitating effective energy management. However, existing methods face challenges such as incomplete information, uncertain parameters, and dynamic environments, which hinder their real-world implementation. This paper proposes an adaptive power capping framework tailored to cloud data centers. By dynamically setting the energy consumption upper bound, the power load of data centers can be reshaped to align with the electricity price or other market signals. To this end, we formulate the power capping problem as a partially observable Markov decision process. Subsequently, we develop an uncertainty-aware model-based reinforcement learning (MBRL) method to perceive the cloud data center operational environment and optimize power-capping decisions. By incorporating a two-stage uncertainty-aware optimization algorithm into the MBRL, we improve its adaptability to the ever-changing environment. Additionally, we derive the optimality gap of the proposed scheme under finite iterations, ensuring effective decisions under complex and uncertain scenarios. The numerical experiments validate the effectiveness of the proposed method using a cloud data center operational environment simulator built on real-world production traces from Alibaba, which demonstrates its potential as an efficient energy management solution for cloud data centers. △ Less

Submitted 9 August, 2025; originally announced August 2025.

Journal ref: IEEE Trans. Smart Grid, Early Access, pp.1-1, Aug.12, 2025

Understanding the Embedding Models on Hyper-relational Knowledge Graph

Authors: Yubo Wang, Shimin Di, Zhili Wang, Haoyang Li, Fei Teng, Hao Xin, Lei Chen

Abstract: Recently, Hyper-relational Knowledge Graphs (HKGs) have been proposed as an extension of traditional Knowledge Graphs (KGs) to better represent real-world facts with additional qualifiers. As a result, researchers have attempted to adapt classical Knowledge Graph Embedding (KGE) models for HKGs by designing extra qualifier processing modules. However, it remains unclear whether the superior perfor… ▽ More Recently, Hyper-relational Knowledge Graphs (HKGs) have been proposed as an extension of traditional Knowledge Graphs (KGs) to better represent real-world facts with additional qualifiers. As a result, researchers have attempted to adapt classical Knowledge Graph Embedding (KGE) models for HKGs by designing extra qualifier processing modules. However, it remains unclear whether the superior performance of Hyper-relational KGE (HKGE) models arises from their base KGE model or the specially designed extension module. Hence, in this paper, we data-wise convert HKGs to KG format using three decomposition methods and then evaluate the performance of several classical KGE models on HKGs. Our results show that some KGE models achieve performance comparable to that of HKGE models. Upon further analysis, we find that the decomposition methods alter the original HKG topology and fail to fully preserve HKG information. Moreover, we observe that current HKGE models are either insufficient in capturing the graph's long-range dependency or struggle to integrate main-triple and qualifier information due to the information compression issue. To further justify our findings and offer a potential direction for future HKGE research, we propose the FormerGNN framework. This framework employs a qualifier integrator to preserve the original HKG topology, and a GNN-based graph encoder to capture the graph's long-range dependencies, followed by an improved approach for integrating main-triple and qualifier information to mitigate compression issues. Our experimental results demonstrate that FormerGNN outperforms existing HKGE models. △ Less

Submitted 5 August, 2025; originally announced August 2025.

Comments: Accepted by CIKM 2025

QuaDreamer: Controllable Panoramic Video Generation for Quadruped Robots

Authors: Sheng Wu, Fei Teng, Hao Shi, Qi Jiang, Kai Luo, Kaiwei Wang, Kailun Yang

Abstract: Panoramic cameras, capturing comprehensive 360-degree environmental data, are suitable for quadruped robots in surrounding perception and interaction with complex environments. However, the scarcity of high-quality panoramic training data-caused by inherent kinematic constraints and complex sensor calibration challenges-fundamentally limits the development of robust perception systems tailored to… ▽ More Panoramic cameras, capturing comprehensive 360-degree environmental data, are suitable for quadruped robots in surrounding perception and interaction with complex environments. However, the scarcity of high-quality panoramic training data-caused by inherent kinematic constraints and complex sensor calibration challenges-fundamentally limits the development of robust perception systems tailored to these embodied platforms. To address this issue, we propose QuaDreamer-the first panoramic data generation engine specifically designed for quadruped robots. QuaDreamer focuses on mimicking the motion paradigm of quadruped robots to generate highly controllable, realistic panoramic videos, providing a data source for downstream tasks. Specifically, to effectively capture the unique vertical vibration characteristics exhibited during quadruped locomotion, we introduce Vertical Jitter Encoding (VJE). VJE extracts controllable vertical signals through frequency-domain feature filtering and provides high-quality prompts. To facilitate high-quality panoramic video generation under jitter signal control, we propose a Scene-Object Controller (SOC) that effectively manages object motion and boosts background jitter control through the attention mechanism. To address panoramic distortions in wide-FoV video generation, we propose the Panoramic Enhancer (PE)-a dual-stream architecture that synergizes frequency-texture refinement for local detail enhancement with spatial-structure correction for global geometric consistency. We further demonstrate that the generated video sequences can serve as training data for the quadruped robot's panoramic visual perception model, enhancing the performance of multi-object tracking in 360-degree scenes. The source code and model weights will be publicly available at https://github.com/losehu/QuaDreamer. △ Less

Submitted 15 October, 2025; v1 submitted 4 August, 2025; originally announced August 2025.

Comments: Accepted to CoRL 2025. The source code and model weights will be publicly available at https://github.com/losehu/QuaDreamer

Two-Stage TSO-DSO Services Provision Framework for Electric Vehicle Coordination

Authors: Yi Wang, Dawei Qiu, Fei Teng, Goran Strbac

Abstract: High renewable penetration has been witnessed in power systems, resulting in reduced system inertia and increasing requirements for frequency response services. Electric vehicles (EVs), owing to their vehicle-to-grid (V2G) capabilities, can provide cost-effective frequency services for transmission system operators (TSOs). However, EVs that are inherently connected to distribution networks may pos… ▽ More High renewable penetration has been witnessed in power systems, resulting in reduced system inertia and increasing requirements for frequency response services. Electric vehicles (EVs), owing to their vehicle-to-grid (V2G) capabilities, can provide cost-effective frequency services for transmission system operators (TSOs). However, EVs that are inherently connected to distribution networks may pose voltage security issues for distribution system operators (DSOs) when supporting TSO frequency. To coordinate both TSO frequency and DSO voltage, this paper proposes a two-stage service provision framework for multi-EVs. At stage one, EVs participate in day-ahead TSO-DSO interactions for frequency reserve schedules; at stage two, EVs make real-time dispatching behaviors in distribution networks for reserve delivery while supporting DSO voltage. Considering the potentially large EV number and environment complexity, a decentralized operation paradigm is introduced for real-time EV dispatches at stage two, while a communication-efficient reinforcement learning (RL) algorithm is proposed to reduce the communication overhead during large-scale multi-agent RL training without compromising policy performance. Case studies are carried out on a 6-bus transmission and 33-bus distribution network as well as a 69-bus distribution network to evaluate the effectiveness and scalability of the proposed method in enabling EVs for frequency service and voltage support. △ Less

Submitted 24 July, 2025; originally announced July 2025.

Towards Microgrid Resilience Enhancement via Mobile Power Sources and Repair Crews: A Multi-Agent Reinforcement Learning Approach

Authors: Yi Wang, Dawei Qiu, Fei Teng, Goran Strbac

Abstract: Mobile power sources (MPSs) have been gradually deployed in microgrids as critical resources to coordinate with repair crews (RCs) towards resilience enhancement owing to their flexibility and mobility in handling the complex coupled power-transport systems. However, previous work solves the coordinated dispatch problem of MPSs and RCs in a centralized manner with the assumption that the communica… ▽ More Mobile power sources (MPSs) have been gradually deployed in microgrids as critical resources to coordinate with repair crews (RCs) towards resilience enhancement owing to their flexibility and mobility in handling the complex coupled power-transport systems. However, previous work solves the coordinated dispatch problem of MPSs and RCs in a centralized manner with the assumption that the communication network is still fully functioning after the event. However, there is growing evidence that certain extreme events will damage or degrade communication infrastructure, which makes centralized decision making impractical. To fill this gap, this paper formulates the resilience-driven dispatch problem of MPSs and RCs in a decentralized framework. To solve this problem, a hierarchical multi-agent reinforcement learning method featuring a two-level framework is proposed, where the high-level action is used to switch decision-making between power and transport networks, and the low-level action constructed via a hybrid policy is used to compute continuous scheduling and discrete routing decisions in power and transport networks, respectively. The proposed method also uses an embedded function encapsulating system dynamics to enhance learning stability and scalability. Case studies based on IEEE 33-bus and 69-bus power networks are conducted to validate the effectiveness of the proposed method in load restoration. △ Less

Submitted 24 July, 2025; originally announced July 2025.

Hallucinating 360°: Panoramic Street-View Generation via Local Scenes Diffusion and Probabilistic Prompting

Authors: Fei Teng, Kai Luo, Sheng Wu, Siyu Li, Pujun Guo, Jiale Wei, Jiaming Zhang, Kunyu Peng, Kailun Yang

Abstract: Panoramic perception holds significant potential for autonomous driving, enabling vehicles to acquire a comprehensive 360° surround view in a single shot. However, autonomous driving is a data-driven task. Complete panoramic data acquisition requires complex sampling systems and annotation pipelines, which are time-consuming and labor-intensive. Although existing street view generation models have… ▽ More Panoramic perception holds significant potential for autonomous driving, enabling vehicles to acquire a comprehensive 360° surround view in a single shot. However, autonomous driving is a data-driven task. Complete panoramic data acquisition requires complex sampling systems and annotation pipelines, which are time-consuming and labor-intensive. Although existing street view generation models have demonstrated strong data regeneration capabilities, they can only learn from the fixed data distribution of existing datasets and cannot leverage stitched pinhole images as a supervisory signal. In this paper, we propose the first panoramic generation method Percep360 for autonomous driving. Percep360 enables coherent generation of panoramic data with control signals based on the stitched panoramic data. Percep360 focuses on two key aspects: coherence and controllability. Specifically, to overcome the inherent information loss caused by the pinhole sampling process, we propose the Local Scenes Diffusion Method (LSDM). LSDM reformulates the panorama generation as a spatially continuous diffusion process, bridging the gaps between different data distributions. Additionally, to achieve the controllable generation of panoramic images, we propose a Probabilistic Prompting Method (PPM). PPM dynamically selects the most relevant control cues, enabling controllable panoramic image generation. We evaluate the effectiveness of the generated images from three perspectives: image quality assessment (i.e., no-reference and with reference), controllability, and their utility in real-world Bird's Eye View (BEV) segmentation. Notably, the generated data consistently outperforms the original stitched images in no-reference quality metrics and enhances downstream perception models. The source code will be publicly available at https://github.com/FeiT-FeiTeng/Percep360. △ Less

Submitted 13 February, 2026; v1 submitted 9 July, 2025; originally announced July 2025.

Comments: Accepted to ICRA 2026. The source code will be publicly available at https://github.com/FeiT-FeiTeng/Percep360

Exploring Partial Multi-Label Learning via Integrating Semantic Co-occurrence Knowledge

Authors: Xin Wu, Fei Teng, Yue Feng, Kaibo Shi, Zhuosheng Lin, Ji Zhang, James Wang

Abstract: Partial multi-label learning aims to extract knowledge from incompletely annotated data, which includes known correct labels, known incorrect labels, and unknown labels. The core challenge lies in accurately identifying the ambiguous relationships between labels and instances. In this paper, we emphasize that matching co-occurrence patterns between labels and instances is key to addressing this ch… ▽ More Partial multi-label learning aims to extract knowledge from incompletely annotated data, which includes known correct labels, known incorrect labels, and unknown labels. The core challenge lies in accurately identifying the ambiguous relationships between labels and instances. In this paper, we emphasize that matching co-occurrence patterns between labels and instances is key to addressing this challenge. To this end, we propose Semantic Co-occurrence Insight Network (SCINet), a novel and effective framework for partial multi-label learning. Specifically, SCINet introduces a bi-dominant prompter module, which leverages an off-the-shelf multimodal model to capture text-image correlations and enhance semantic alignment. To reinforce instance-label interdependencies, we develop a cross-modality fusion module that jointly models inter-label correlations, inter-instance relationships, and co-occurrence patterns across instance-label assignments. Moreover, we propose an intrinsic semantic augmentation strategy that enhances the model's understanding of intrinsic data semantics by applying diverse image transformations, thereby fostering a synergistic relationship between label confidence and sample difficulty. Extensive experiments on four widely-used benchmark datasets demonstrate that SCINet surpasses state-of-the-art methods. △ Less

Submitted 22 February, 2026; v1 submitted 8 July, 2025; originally announced July 2025.

Comments: Accepted by IEEE Transactions on Multimedia

BlueLM-2.5-3B Technical Report

Authors: Baojiao Xiong, Boheng Chen, Chengzhi Wang, Daxiong Luo, Dongsheng Xu, Dongyang Liu, Fan Yang, Fangyuan Li, Fei Teng, Feng Wang, Fukang Qin, Fuquan Peng, Guanxin Tan, Guozhi Wang, Haibo Yu, Haohao Gao, Heng Liu, Hongbo Yang, Hongjian Zou, Houzheng Shen, Hu Meng, Huan Li, Hui Tan, Jiali Chen, Jianzhao Chen , et al. (36 additional authors not shown)

Abstract: We present BlueLM-2.5-3B, a compact and unified dense Multimodal Large Language Model (MLLM) designed for efficient edge-device deployment, offering strong general-purpose and reasoning capabilities. To the best of our knowledge, this is the first 3B-scale MLLM to support both thinking and non-thinking modes, while also enabling explicit control over thinking token budget. BlueLM-2.5-3B is develop… ▽ More We present BlueLM-2.5-3B, a compact and unified dense Multimodal Large Language Model (MLLM) designed for efficient edge-device deployment, offering strong general-purpose and reasoning capabilities. To the best of our knowledge, this is the first 3B-scale MLLM to support both thinking and non-thinking modes, while also enabling explicit control over thinking token budget. BlueLM-2.5-3B is developed through diversified data curation, key data resampling, hybrid heterogeneous reinforcement learning, and a high-performance training infrastructure. Our model achieves superior multimodal capacity while preserving competitive pure-text performance with only 2.9 billion parameters. We conduct comprehensive evaluations across a broad range of multimodal and text-only benchmarks. In thinking mode, BlueLM-2.5-3B achieves comparable performance to Qwen3-4B on text-only benchmarks, and trails the larger Kimi-VL-A3B-16B by only about 5% on average across multimodal evaluations. In non-thinking mode, it outperforms Qwen2.5-VL-3B on the majority of multimodal benchmarks. Additionally, BlueLM-2.5-3B exhibits exceptional data efficiency. All of the aforementioned performance is achieved with substantially less total training data than Qwen2.5-VL-3B and Qwen3-4B. We hope our work contributes to the advancement of high-performance, on-device MLLMs and provides meaningful insights to the research community. △ Less

Submitted 8 July, 2025; originally announced July 2025.

NRSeg: Noise-Resilient Learning for BEV Semantic Segmentation via Driving World Models

Authors: Siyu Li, Fei Teng, Yihong Cao, Kailun Yang, Zhiyong Li, Yaonan Wang

Abstract: Birds' Eye View (BEV) semantic segmentation is an indispensable perception task in end-to-end autonomous driving systems. Unsupervised and semi-supervised learning for BEV tasks, as pivotal for real-world applications, underperform due to the homogeneous distribution of the labeled data. In this work, we explore the potential of synthetic data from driving world models to enhance the diversity of… ▽ More Birds' Eye View (BEV) semantic segmentation is an indispensable perception task in end-to-end autonomous driving systems. Unsupervised and semi-supervised learning for BEV tasks, as pivotal for real-world applications, underperform due to the homogeneous distribution of the labeled data. In this work, we explore the potential of synthetic data from driving world models to enhance the diversity of labeled data for robustifying BEV segmentation. Yet, our preliminary findings reveal that generation noise in synthetic data compromises efficient BEV model learning. To fully harness the potential of synthetic data from world models, this paper proposes NRSeg, a noise-resilient learning framework for BEV semantic segmentation. Specifically, a Perspective-Geometry Consistency Metric (PGCM) is proposed to quantitatively evaluate the guidance capability of generated data for model learning. This metric originates from the alignment measure between the perspective road mask of generated data and the mask projected from the BEV labels. Moreover, a Bi-Distribution Parallel Prediction (BiDPP) is designed to enhance the inherent robustness of the model, where the learning process is constrained through parallel prediction of multinomial and Dirichlet distributions. The former efficiently predicts semantic probabilities, whereas the latter adopts evidential deep learning to realize uncertainty quantification. Furthermore, a Hierarchical Local Semantic Exclusion (HLSE) module is designed to address the non-mutual exclusivity inherent in BEV semantic segmentation tasks. Experimental results demonstrate that NRSeg achieves state-of-the-art performance, yielding the highest improvements in mIoU of 13.8% and 11.4% in unsupervised and semi-supervised BEV segmentation tasks, respectively. The source code will be made publicly available at https://github.com/lynn-yu/NRSeg. △ Less

Submitted 24 February, 2026; v1 submitted 5 July, 2025; originally announced July 2025.

Comments: Accepted to IEEE Transactions on Image Processing (TIP). The source code will be made publicly available at https://github.com/lynn-yu/NRSeg

Exploiting Vision Language Model for Training-Free 3D Point Cloud OOD Detection via Graph Score Propagation

Authors: Tiankai Chen, Yushu Li, Adam Goodge, Fei Teng, Xulei Yang, Tianrui Li, Xun Xu

Abstract: Out-of-distribution (OOD) detection in 3D point cloud data remains a challenge, particularly in applications where safe and robust perception is critical. While existing OOD detection methods have shown progress for 2D image data, extending these to 3D environments involves unique obstacles. This paper introduces a training-free framework that leverages Vision-Language Models (VLMs) for effective… ▽ More Out-of-distribution (OOD) detection in 3D point cloud data remains a challenge, particularly in applications where safe and robust perception is critical. While existing OOD detection methods have shown progress for 2D image data, extending these to 3D environments involves unique obstacles. This paper introduces a training-free framework that leverages Vision-Language Models (VLMs) for effective OOD detection in 3D point clouds. By constructing a graph based on class prototypes and testing data, we exploit the data manifold structure to enhancing the effectiveness of VLMs for 3D OOD detection. We propose a novel Graph Score Propagation (GSP) method that incorporates prompt clustering and self-training negative prompting to improve OOD scoring with VLM. Our method is also adaptable to few-shot scenarios, providing options for practical applications. We demonstrate that GSP consistently outperforms state-of-the-art methods across synthetic and real-world datasets 3D point cloud OOD detection. △ Less

Submitted 27 June, 2025; originally announced June 2025.

Comments: Accepted by ICCV 2025

Out-of-Distribution Semantic Occupancy Prediction

Authors: Yuheng Zhang, Mengfei Duan, Kunyu Peng, Yuhang Wang, Ruiping Liu, Fei Teng, Kai Luo, Zhiyong Li, Kailun Yang

Abstract: 3D semantic occupancy prediction is crucial for autonomous driving, providing a dense, semantically rich environmental representation. However, existing methods focus on in-distribution scenes, making them susceptible to Out-of-Distribution (OoD) objects and long-tail distributions, which increases the risk of undetected anomalies and misinterpretations, posing safety hazards. To address these cha… ▽ More 3D semantic occupancy prediction is crucial for autonomous driving, providing a dense, semantically rich environmental representation. However, existing methods focus on in-distribution scenes, making them susceptible to Out-of-Distribution (OoD) objects and long-tail distributions, which increases the risk of undetected anomalies and misinterpretations, posing safety hazards. To address these challenges, we introduce Out-of-Distribution Semantic Occupancy Prediction, targeting OoD detection in 3D voxel space. To fill dataset gaps, we propose a Realistic Anomaly Augmentation that injects synthetic anomalies while preserving realistic spatial and occlusion patterns, enabling the creation of two datasets: VAA-KITTI and VAA-KITTI-360. Then, a novel framework that integrates OoD detection into 3D semantic occupancy prediction, OccOoD, is proposed, which uses Cross-Space Semantic Refinement (CSSR) to refine semantic predictions from complementary voxel and BEV representations, improving OoD detection. Experimental results demonstrate that OccOoD achieves state-of-the-art OoD detection with an AuROC of 65.50% and an AuPRCr of 31.83 within a 1.2m region, while maintaining competitive semantic occupancy prediction performance and generalization in real-world urban driving scenes. The established datasets and source code will be made publicly available at https://github.com/7uHeng/OccOoD. △ Less

Submitted 10 January, 2026; v1 submitted 26 June, 2025; originally announced June 2025.

Comments: The established datasets and source code will be made publicly available at https://github.com/7uHeng/OccOoD

Go Beyond Earth: Understanding Human Actions and Scenes in Microgravity Environments

Authors: Di Wen, Lei Qi, Kunyu Peng, Kailun Yang, Fei Teng, Ao Luo, Jia Fu, Yufan Chen, Ruiping Liu, Yitian Shi, M. Saquib Sarfraz, Rainer Stiefelhagen

Abstract: Despite substantial progress in video understanding, most existing datasets are limited to Earth's gravitational conditions. However, microgravity alters human motion, interactions, and visual semantics, revealing a critical gap for real-world vision systems. This presents a challenge for domain-robust video understanding in safety-critical space applications. To address this, we introduce MicroG-… ▽ More Despite substantial progress in video understanding, most existing datasets are limited to Earth's gravitational conditions. However, microgravity alters human motion, interactions, and visual semantics, revealing a critical gap for real-world vision systems. This presents a challenge for domain-robust video understanding in safety-critical space applications. To address this, we introduce MicroG-4M, the first benchmark for spatio-temporal and semantic understanding of human activities in microgravity. Constructed from real-world space missions and cinematic simulations, the dataset includes 4,759 clips covering 50 actions, 1,238 context-rich captions, and over 7,000 question-answer pairs on astronaut activities and scene understanding. MicroG-4M supports three core tasks: fine-grained multi-label action recognition, temporal video captioning, and visual question answering, enabling a comprehensive evaluation of both spatial localization and semantic reasoning in microgravity contexts. We establish baselines using state-of-the-art models. All data, annotations, and code are available at https://github.com/LEI-QI-233/HAR-in-Space. △ Less

Submitted 22 March, 2026; v1 submitted 3 June, 2025; originally announced June 2025.

Comments: 16 pages, 4 figures, code are available at https://github.com/LEI-QI-233/HAR-in-Space

Synergising Hierarchical Data Centers and Power Networks: A Privacy-Preserving Approach

Authors: Junhong Liu, Fei Teng, Yunhe Hou

Abstract: In the era of digitization, data centers have emerged as integral contributors sustaining our interlinked world, bearing responsibility for an increasing proportion of the world's energy consumption. To facilitate the their fast rollout while progressing towards net-zero energy systems, the synergy of hierarchical data centers (cloud-fog-edge) and power networks can play a pivotal role. However, e… ▽ More In the era of digitization, data centers have emerged as integral contributors sustaining our interlinked world, bearing responsibility for an increasing proportion of the world's energy consumption. To facilitate the their fast rollout while progressing towards net-zero energy systems, the synergy of hierarchical data centers (cloud-fog-edge) and power networks can play a pivotal role. However, existing centralized co-dispatch manners encroach on the privacy of different agents within the integrated systems, meanwhile suffering from the combinatorial explosion. In this research, we propose a near-optimal distributed privacy-preserving approach to solve the non-convex synergy (day-ahead co-dispatch) problem. The synergy problem is formulated as a mixed integer quadratically constrained quadratic programming considering both communication and energy conservation, where Lyapunov optimization is introduced to balance operating costs and uncertain communication delays. To mitigate impacts of the highly non-convex nature, the normalized multi-parametric disaggregation technique is leveraged to reformulate the problem into a mixed integer non-linear programming. To further overcome non-smoothness of the reformulated problem, the customized $\ell_1-$surrogate Lagrangian relaxation method with convergence guarantees is proposed to solve the problem in a distributed privacy-preserving manner. The effectiveness, optimality, and scalability of the proposed methodologies for the synergy problem are validated via numerical simulations. Simulation results also indicate that computing tasks can be delayed and migrated within the hierarchical data centers, demonstrating the flexible resource allocation capabilities of the hierarchical data center architecture, further facilitating peak load balancing in the power network. △ Less

Submitted 24 August, 2025; v1 submitted 26 May, 2025; originally announced May 2025.

Report number: Volume: 16, Issue: 6, November 2025

Journal ref: IEEE Transactions on Smart Grid, Volume: 16, Issue: 6, November 2025

Learning-Augmented Power System Operations: A Unified Optimization View

Authors: Wangkun Xu, Zhongda Chu, Fei Teng

Abstract: With the increasing penetration of renewable energy and inverter-based resources, traditional physics-based power-system operation faces growing challenges in maintaining economic efficiency, security, and robustness. Machine learning (ML) has emerged as a powerful tool for modeling complex system dynamics and uncertainty. However, standalone ML pipelines, including model selection, training, and… ▽ More With the increasing penetration of renewable energy and inverter-based resources, traditional physics-based power-system operation faces growing challenges in maintaining economic efficiency, security, and robustness. Machine learning (ML) has emerged as a powerful tool for modeling complex system dynamics and uncertainty. However, standalone ML pipelines, including model selection, training, and validation, are often designed separately from the downstream optimization problems they influence, which can lead to suboptimal system-level decisions. To address this gap, this paper proposes \emph{Learning-Augmented Power System Operations} (LAPSO), a unified optimization-centered framework that treats ML as an explicit component of power-system operational decision-making. First, LAPSO provides generalized mathematical template covering both decision-independent predictors that parameterize downstream optimization and decision-dependent learned surrogates that enter optimization as auxiliary constraints. Second, it designs ML pipelines using optimization-aware criteria, including solution-quality, computational tractability, constraint satisfaction, and economic performance. We instantiate LAPSO on both stability-constrained optimization (SCO) and objective-based forecasting (OBF), and show how the framework provides actionable guidance for selecting learned components. We further extend the framework to a hybrid forecast--operation--control chain and use it to organize heterogeneous uncertainty sources. Finally, we release an open-source Python package, \texttt{lapso}, for modularly augmenting existing power-system optimization models with ML components. Code and datasets are available at: https://github.com/xuwkk/lapso_exp. △ Less

Submitted 19 August, 2026; v1 submitted 8 May, 2025; originally announced May 2025.

Panoramic Out-of-Distribution Segmentation

Authors: Mengfei Duan, Yuheng Zhang, Yihong Cao, Fei Teng, Kai Luo, Jiaming Zhang, Kailun Yang, Zhiyong Li

Abstract: Panoramic imaging enables capturing 360° images with an ultra-wide Field-of-View (FoV) for dense omnidirectional perception, which is critical to applications, such as autonomous driving and augmented reality, etc. However, current panoramic semantic segmentation methods fail to identify outliers, and pinhole Out-of-distribution Segmentation (OoS) models perform unsatisfactorily in the panoramic d… ▽ More Panoramic imaging enables capturing 360° images with an ultra-wide Field-of-View (FoV) for dense omnidirectional perception, which is critical to applications, such as autonomous driving and augmented reality, etc. However, current panoramic semantic segmentation methods fail to identify outliers, and pinhole Out-of-distribution Segmentation (OoS) models perform unsatisfactorily in the panoramic domain due to pixel distortions and background clutter. To address these issues, we introduce a new task, Panoramic Out-of-distribution Segmentation (PanOoS), with the aim of achieving comprehensive and safe scene understanding. Furthermore, we propose the first solution, POS, which adapts to the characteristics of panoramic images through text-guided prompt distribution learning. Specifically, POS integrates a disentanglement strategy designed to materialize the cross-domain generalization capability of CLIP. The proposed Prompt-based Restoration Attention (PRA) optimizes semantic decoding by prompt guidance and self-adaptive correction, while Bilevel Prompt Distribution Learning (BPDL) refines the manifold of per-pixel mask embeddings via semantic prototype supervision. Besides, to compensate for the scarcity of PanOoS datasets, we establish two benchmarks: DenseOoS, which features diverse outliers in complex environments, and QuadOoS, captured by a quadruped robot with a panoramic annular lens system. Extensive experiments demonstrate superior performance of POS, with AuPRC improving by 34.25% and FPR95 decreasing by 21.42% on DenseOoS, outperforming state-of-the-art pinhole-OoS methods. Moreover, POS achieves leading closed-set segmentation capabilities and advances the development of panoramic understanding. Code and datasets will be available at https://github.com/MengfeiD/PanOoS. △ Less

Submitted 11 December, 2025; v1 submitted 6 May, 2025; originally announced May 2025.

Comments: Code and datasets will be available at https://github.com/MengfeiD/PanOoS

Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systems

Authors: Bang Liu, Xinfeng Li, Jiayi Zhang, Jinlin Wang, Tanjin He, Sirui Hong, Hongzhang Liu, Shaokun Zhang, Kaitao Song, Kunlun Zhu, Yuheng Cheng, Suyuchen Wang, Xiaoqiang Wang, Yuyu Luo, Haibo Jin, Peiyan Zhang, Ollie Liu, Jiaqi Chen, Huan Zhang, Zhaoyang Yu, Haochen Shi, Boyan Li, Dekun Wu, Fengwei Teng, Xiaojun Jia , et al. (23 additional authors not shown)

Abstract: The advent of large language models (LLMs) has catalyzed a transformative shift in artificial intelligence, paving the way for advanced intelligent agents capable of sophisticated reasoning, robust perception, and versatile action across diverse domains. As these agents increasingly drive AI research and practical applications, their design, evaluation, and continuous improvement present intricate… ▽ More The advent of large language models (LLMs) has catalyzed a transformative shift in artificial intelligence, paving the way for advanced intelligent agents capable of sophisticated reasoning, robust perception, and versatile action across diverse domains. As these agents increasingly drive AI research and practical applications, their design, evaluation, and continuous improvement present intricate, multifaceted challenges. This book provides a comprehensive overview, framing intelligent agents within modular, brain-inspired architectures that integrate principles from cognitive science, neuroscience, and computational research. We structure our exploration into four interconnected parts. First, we systematically investigate the modular foundation of intelligent agents, systematically mapping their cognitive, perceptual, and operational modules onto analogous human brain functionalities and elucidating core components such as memory, world modeling, reward processing, goal, and emotion. Second, we discuss self-enhancement and adaptive evolution mechanisms, exploring how agents autonomously refine their capabilities, adapt to dynamic environments, and achieve continual learning through automated optimization paradigms. Third, we examine multi-agent systems, investigating the collective intelligence emerging from agent interactions, cooperation, and societal structures. Finally, we address the critical imperative of building safe and beneficial AI systems, emphasizing intrinsic and extrinsic security threats, ethical alignment, robustness, and practical mitigation strategies necessary for trustworthy real-world deployment. By synthesizing modular AI architectures with insights from different disciplines, this survey identifies key research challenges and opportunities, encouraging innovations that harmonize technological advancement with meaningful societal benefit. △ Less

Submitted 2 August, 2025; v1 submitted 31 March, 2025; originally announced April 2025.