Showing 1–50 of 87 results for author: Hua, J

E-S2Feat:Semantic-Guided Spiking Local Feature Detection and Description for Event Cameras

Authors: Yang Yi, Juntao Hua, Jinpu Zhang, Liangwei Fan, Hui Shen, Dewen Hu

Abstract: Benefiting from high temporal resolution and dynamic range, event-based local feature methods have attracted increasing attention. However, event sparsity, noise, and limited texture still hinder robust local feature learning. Deploying such methods on resource-constrained platforms such as unmanned aerial vehicles also requires balancing accuracy and energy efficiency. To address these challenges… ▽ More Benefiting from high temporal resolution and dynamic range, event-based local feature methods have attracted increasing attention. However, event sparsity, noise, and limited texture still hinder robust local feature learning. Deploying such methods on resource-constrained platforms such as unmanned aerial vehicles also requires balancing accuracy and energy efficiency. To address these challenges, this paper proposes \textbf{E-S2Feat}, a spiking neural network framework for event-based local feature detection and description. The framework jointly optimizes local feature learning from the perspectives of feature representation and selection. First, a module-specific spiking activation mechanism preserves fine-grained structural cues and discriminative information under low-bit, energy-efficient inference, thereby improving overall representation fidelity. Furthermore, a semantic-guided feature modulation mechanism leverages semantic priors to refine keypoint response distributions and enhance local descriptor discriminability, thereby guiding the model to extract local features with greater geometric stability and stronger discriminative capability. Experiments on the ECD and EDS datasets show that the proposed method significantly outperforms baseline methods such as SuperEvent in pose estimation accuracy. It also achieves accuracy comparable to its artificial neural network counterpart while delivering an approximately 4.8-fold improvement in theoretical computational energy efficiency. Visual-inertial odometry experiments on the TUM-VIE dataset further verify the effectiveness and practical application potential of the proposed method in complete SLAM systems. △ Less

Submitted 14 August, 2026; originally announced August 2026.

PURPOSE: Poisoning Conflict Resolution in RAG via Proxy-Fact-Grounded Updates

Authors: Zijian Wang, Yubo Zhu, Muzhi Dong, Yanjun Lou, Yisheng Li, ZiLiang Zhang, Wei Tong, Yuan Zhang, Jingyu Hua, Sheng Zhong

Abstract: In Retrieval-Augmented Generation (RAG), post-retrieval conflict resolution arbitrates among noisy or contradictory retrieved passages. However, the robustness of this safeguard against knowledge poisoning has not been adequately studied. Existing black-box poisoning methods all assert the target answer in frontal contradiction with what the resolver treats as settled, the very signal these method… ▽ More In Retrieval-Augmented Generation (RAG), post-retrieval conflict resolution arbitrates among noisy or contradictory retrieved passages. However, the robustness of this safeguard against knowledge poisoning has not been adequately studied. Existing black-box poisoning methods all assert the target answer in frontal contradiction with what the resolver treats as settled, the very signal these methods are built to detect. We propose PURPOSE, a strict black-box poisoning attack that reframes the injection as an update that minimizes conflict, rather than as a counter-claim. PURPOSE extracts query-related facts approximating the resolver's possible reference, then grounds a pivot event in them to keep the injection consistent with what the resolver might verify while steering the generator toward the target answer. Across three QA benchmarks, five generators, and three conflict-resolution methods, PURPOSE attains the highest attack success rate (ASR) in 35 of 45 settings and exceeds the strongest prior attack with +9.7 mean ASR points. These results show that our poisoning method is effective against conflict resolution in RAG and identify non-contradicting injection as a practical mode to enhance poisoning attack. △ Less

Submitted 5 August, 2026; originally announced August 2026.

CAPE-T2V: Captioner-Anchored Prompt Enhancement toward Two-Sided Conditioning Alignment in Text-to-Video Generation

Authors: Yizhuo Jia, Jingyun Hua, Yuanxing Zhang

Abstract: Text-to-video (T2V) diffusion transformers (DiTs) are trained with detailed video captions, whereas inference often relies on user prompts rewritten by a prompt enhancer (PE). Prior work has improved generation by optimizing the PE, the DiT, or both; some methods have also sought to narrow the training-inference mismatch through shared schemas. Yet even within a shared schema, inference-time PE ou… ▽ More Text-to-video (T2V) diffusion transformers (DiTs) are trained with detailed video captions, whereas inference often relies on user prompts rewritten by a prompt enhancer (PE). Prior work has improved generation by optimizing the PE, the DiT, or both; some methods have also sought to narrow the training-inference mismatch through shared schemas. Yet even within a shared schema, inference-time PE outputs and DiT training captions may still differ in detail selection, information organization, descriptive granularity, and phrasing. We refer to this residual mismatch as the PE-Caption gap and introduce CAPE-T2V, a two-step Captioner-Anchored Prompt Enhancement framework toward two-sided conditioning alignment in T2V generation. First, CAPE-T2V constructs three types of PE training examples, pairing captioner-generated targets with concise source captions, detailed source captions, or pseudo user prompts derived from those targets. It then fine-tunes the PE to map each input to its paired target. Second, CAPE-T2V fine-tunes the DiT on video-derived captions rewritten by the Anchored PE; the same PE rewrites user prompts at inference. Relative to a baseline using the same caption schema, CAPE-T2V achieves higher aggregate scores on StoryEval, VBench-2.0, and T2V-CompBench across Wan2.2 and LTX-2.3. Further, CAPE-T2V exhibits a smaller PE-Caption gap than the baseline: its DiT fine-tuning captions are closer in distribution to inference-time PE outputs, as measured by squared maximum mean discrepancy in a fixed embedding space. Overall, these results support CAPE-T2V as an effective approach to mitigating the PE-Caption gap. The project is available at https://github.com/yizzz927/CAPE-T2V. △ Less

Submitted 3 August, 2026; originally announced August 2026.

Comments: Includes appendix; 11 figures. Project page: https://github.com/yizzz927/CAPE-T2V

VC-Tooler: Learning Compositional and Adaptive Visual Tool Use

Authors: Yizheng Wu, Jiashen Hua, Bing Deng, Jieping Ye

Abstract: Agentic multimodal reasoning extends passive image understanding by allowing VLMs to actively acquire and refine visual evidence through visual tool interactions. Effective visual tool use requires three capabilities: grounding tool calls in visual context, composing tools across multiple steps, and adapting reasoning to tool-returned observations. However, existing approaches largely focus on gro… ▽ More Agentic multimodal reasoning extends passive image understanding by allowing VLMs to actively acquire and refine visual evidence through visual tool interactions. Effective visual tool use requires three capabilities: grounding tool calls in visual context, composing tools across multiple steps, and adapting reasoning to tool-returned observations. However, existing approaches largely focus on grounding within fixed tool spaces and rigid invocation patterns, leaving composition and adaptation insufficiently addressed. We present VC-Tooler, which learns visual tool use as a compositional and adaptive capability. To this end, we first build a trajectory bank through a hierarchical synthesis pipeline covering three capability levels: single-tool grounding, multi-tool composition, and diverse tool contexts and interfaces. We then train the model in two stages: a supervised cold start that establishes these capabilities, followed by reinforcement learning that encourages accurate, efficient, and context-aware visual tool use. VC-Tooler achieves state-of-the-art performance among open-source models on both general-purpose and agentic benchmarks, including $95.8\%$ on V* and $35.3\%$ on VTC-Bench, and shows promising transfer under richer tool settings at inference time. Project page: https://w1zheng.github.io/VC-Tooler △ Less

Submitted 3 August, 2026; originally announced August 2026.

Illuminating Visual Identity in Universal Multimodal Embeddings

Authors: Jiawei Cao, Junyi Feng, Jiashen Hua, Ziheng Huang, Bing Deng, Kaijie Wu, Chaochen Gu, Jieping Ye

Abstract: Universal Multimodal Embeddings (UMEs) aim to unify various modalities and tasks into a shared representation space. In recent years, this field has witnessed substantial progress driven by the development of Multimodal Large Language Models (MLLMs). However, a crucial capability, visual identity discrimination, remains underexplored in existing UME methods, despite its critical role in a wide ran… ▽ More Universal Multimodal Embeddings (UMEs) aim to unify various modalities and tasks into a shared representation space. In recent years, this field has witnessed substantial progress driven by the development of Multimodal Large Language Models (MLLMs). However, a crucial capability, visual identity discrimination, remains underexplored in existing UME methods, despite its critical role in a wide range of tasks, including instance retrieval, re-identification, and identity preservation in AI-generated content. To bridge this gap, we propose a unified formulation for visual identity discrimination~(VisID) and introduce $\textbf{MVEB}$ ($\textbf{M}$ultimodal $\textbf{V}$isual Identity $\textbf{E}$mbedding $\textbf{B}$enchmark), a large-scale benchmark curated from both real-world and synthetic datasets to support evaluation and training. Furthermore, we present a simple yet effective learning framework that jointly optimizes general multimodal and visual identity representations through a carefully designed identity-aware sampling mechanism. Extensive experiments demonstrate that our approach successfully endows UMEs with strong identity discrimination capability and maintains competitive general multimodal performance. We believe this work not only illuminates a critical yet neglected capability, but also takes a step toward more holistic universal multimodal embeddings. Code and data are available at \href{https://chrisclear3.github.io/MVEB}{MVEB}. △ Less

Submitted 3 August, 2026; originally announced August 2026.

Comments: Accepted to CVPR 2026

Internal Algebraic Type Theory

Authors: Joseph Hua

Abstract: This thesis brings us closer to applying computer-assisted, internal, type-theoretic reasoning to a category, with examples in the category of cubical sets, the category of groupoids, and the category of categories. The steps we make towards this general goal are both in furthering the type theoretic analysis of these examples, as well as implementing computer-assisted syntax-semantic reasoning as… ▽ More This thesis brings us closer to applying computer-assisted, internal, type-theoretic reasoning to a category, with examples in the category of cubical sets, the category of groupoids, and the category of categories. The steps we make towards this general goal are both in furthering the type theoretic analysis of these examples, as well as implementing computer-assisted syntax-semantic reasoning as part of the HoTTLean project. One key component of this work is the consideration of new exponentiability conditions with respect to a class of maps in a category, and the development of polynomial functors for these maps, making the methods of "algebraic type theory" possible in this very general setting. △ Less

Submitted 30 July, 2026; originally announced August 2026.

Learning Manifolds in High-D Point Embedding for Anisotropic Surface Approximation from Unstructured Point Clouds

Authors: Hongbo Li, Haikuan Zhu, Xiaohu Guo, Wenping Wang, Jing Hua, Zichun Zhong

Abstract: Dense 3D sensors in various real-world fields produce point clouds that are geometrically redundant for real-time processing. In this paper, we propose an efficient and scalable learning-based anisotropic surface approximation framework, HD-PEA, that operates directly on unstructured point clouds, integrating anisotropic optimization into reconstruction to produce compact, geometry-aligned surface… ▽ More Dense 3D sensors in various real-world fields produce point clouds that are geometrically redundant for real-time processing. In this paper, we propose an efficient and scalable learning-based anisotropic surface approximation framework, HD-PEA, that operates directly on unstructured point clouds, integrating anisotropic optimization into reconstruction to produce compact, geometry-aligned surface representations with higher fidelity, fewer elements, and improved numerical stability compared to isotropic and adaptive meshes. Firstly, we develop a novel learning-based high-dimensional (high-d) Euclidean point embedding method to map the input point clouds into a high-d manifold embedding space. For handling large-scale point clouds without retraining and fine-tuning, a patch-based meta-embedding scheme is designed during the inference stage. Then, we develop a new tangent subspace estimation for the high-d embedding manifold approximation and anisotropic manifold reconstruction in high-d space. The main contribution of this work is to propose a scalable deep learning framework and a variety of datasets for constructing a high-d Euclidean point embedding space aimed to 3D anisotropic surface mesh approximation and Riemannian curvature tensor estimation from point clouds. We extensively evaluate our method against state-of-the-art surface reconstruction approaches using several datasets, such as Thingi10K dataset, AIM@SHAPE and Stanford 3D Scanning Repository, ScanNet dataset, and further demonstrate its generalization and usability on diverse unseen shapes and applications from these datasets. △ Less

Submitted 30 July, 2026; originally announced July 2026.

JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA Models

Authors: Haoran Sun, Wentao Zhang, Junyang Hua, Hedan Yang, Yongjian Guo, Yifei Zhang, Xiaolong Xiang, Mingxi Luo, Jing Long, Chen Zhao, Chen Zhou, Wanting Xu, Qiming Yang, Hui Zhang, Song Wang, Xiaodong Bai, Shuai Di, Xu Chu, Xiaotie Deng, Yicheng Gong, Junwu Xiong

Abstract: The post-training of Vision-Language-Action (VLA) models is essential due to the diversity of simulators, robot embodiments, and task objectives. Existing compute services, whether offered as direct accelerator rental or batch-workload submission, typically allocate an exclusive set of GPU and CPU resources to a single tenant. While this paradigm maximizes client flexibility, it burdens users with… ▽ More The post-training of Vision-Language-Action (VLA) models is essential due to the diversity of simulators, robot embodiments, and task objectives. Existing compute services, whether offered as direct accelerator rental or batch-workload submission, typically allocate an exclusive set of GPU and CPU resources to a single tenant. While this paradigm maximizes client flexibility, it burdens users with infrastructure adaptation, and the fixed card-hour accounting model renders short or bursty workloads both expensive for tenants and inefficient for the service provider. To address these challenges, we present JoyNexus, a unified service for multi-tenant VLA supervised fine-tuning, reinforcement learning, and evaluation. JoyNexus decouples the Training Model Service, Inference Model Service, and Environment Service, each accessed through APIs and backed by resident shared base models with tenant-specific slots. Tenants can directly invoke high-level semantic APIs for training, rollout, and evaluation, or compose custom algorithms using lower-level APIs and their assigned endpoints. Multiple tenants submit workloads concurrently; their action modules, optimizers, rollout records, and policy versions remain isolated, and the service is scheduled by the global Training Queue and Inference Queue. To further improve multi-tenant training efficiency, JoyNexus introduces group batching for heterogeneous VLA data schemas that share a compatible model-facing prefix, enabling a single shared backbone forward pass over grouped samples. Finally, we evaluate JoyNexus through workload simulation and a group-batching pipeline in a realistic embodied scenario. Results show that, compared with isolated single-tenant execution, JoyNexus reduces aggregate GPU time and improves service utilization via cross-tenant scheduling on shared resources. △ Less

Submitted 17 July, 2026; originally announced July 2026.

Comments: 23 pages, 12 figures

LQCDMaster: Agentic Scientific Computing for Lattice Quantum Chromodynamics Research

Authors: Haofei Gao, Tingjia Miao, Wenkai Jin, Muhua Zhang, Hanzhang Wang, Jie Ran, Jinxin Tan, Zhentao Zhang, Bo Tang, Leiyi Li, Jun Hua, Xiangyu Jiang, Qi-An Zhang, Siheng Chen, Wei Wang

Abstract: Lattice quantum chromodynamics (LQCD) provides a first-principles framework for computing hadronic observables, but its practical use remains limited by the substantial expertise required to turn research motivation into reliable computing workflows. Here we present \textsc{LQCDMaster}, a tool-augmented, skill-guided and domain-specialized scientific computing agent that converts natural-language… ▽ More Lattice quantum chromodynamics (LQCD) provides a first-principles framework for computing hadronic observables, but its practical use remains limited by the substantial expertise required to turn research motivation into reliable computing workflows. Here we present \textsc{LQCDMaster}, a tool-augmented, skill-guided and domain-specialized scientific computing agent that converts natural-language LQCD research tasks into executable PyQUDA computing workflows, including measurement scripts, job-submission artifacts, execution logs and numerical outputs. The system combines agentic planning, expert-annotated LQCD skills and a deterministic Wick-contraction tool to constrain the algebraically fragile components of code generation. We evaluate \textsc{LQCDMaster} on a benchmark at the forefront of scientific research, comprising 70 LQCD computing tasks, with observables covering local and nonlocal two-point functions, Wilson loops, meson and baryon three-point functions. The generated workflows exactly reproduce expert-written implementations in 63 of 70 tasks at machine precision, with three additional discrepancies attributable to convention mismatches. Across representative observables, the agent reduces implementation time from hours to minutes while preserving end-to-end numerical validation. Further, we present a typical case of \textsc{LQCDMaster}-driven exploration: a lattice computation of light-cone distribution amplitudes with diagonal Wilson-line, a quantity accessible with standard methods but never before computed, and computation of the spectrum of proton, deuteron, triton, hyperon, hyperdeuteron and hypertriton. This work pioneers the paradigm of agentic scientific computing by automating the end-to-end scientific computing workflows in lattice QCD research, lowering its barrier and facilitating the exploration and verification of non-standard scientific ideas. △ Less

Submitted 16 July, 2026; originally announced July 2026.

Comments: 17 pages, 4 figures

HyperDFlash: Hyper-Connection-Aligned Block Speculative Decoding with Gated Residual Reduction

Authors: Luxi Lin, Shuang Peng, Rui Ma, Junhao Hua, Shuwei Fan, Zhengda Qin, Qiang Wang, Hongjian Sun, Fangmin Chen, Songwei Liu

Abstract: We present HyperDFlash, a block-parallel speculative decoding framework tailored to DeepSeek-V4's Hyper-Connections (HC). Despite the strong performance of DeepSeek-V4's native Multi-Token Prediction (MTP) module on initial token drafting, its draft accuracy degrades sharply at later positions, as error accumulation from unverified intermediate tokens harms draft acceptance rates. Although the ori… ▽ More We present HyperDFlash, a block-parallel speculative decoding framework tailored to DeepSeek-V4's Hyper-Connections (HC). Despite the strong performance of DeepSeek-V4's native Multi-Token Prediction (MTP) module on initial token drafting, its draft accuracy degrades sharply at later positions, as error accumulation from unverified intermediate tokens harms draft acceptance rates. Although the original DFlash method supports efficient one-pass block drafting, it cannot be seamlessly adapted to the HC paradigm, since DeepSeek-V4's multi-path residual stream induces inherent feature misalignment with conventional drafting designs. To resolve this architectural mismatch, we propose two dedicated, model-aligned optimizations for HC residual streams. First, we adopt pre-collapse residual states as the exclusive conditioning signal, preserving complete multi-path structural information and better aligning the drafter with the target's native prediction pathway. Second, we replace the heavy generic linear compressor with a lightweight gated residual reducer, whose parameters are directly inherited from the target model's built-in hc_head module. This design yields input-aware path aggregation with three orders of magnitude fewer parameters while maintaining precise architectural alignment. We further enhance model training via a targeted KL distillation loss applied to the LM-head, regularizing predictions against the target distribution to improve early draft quality. Extensive experiments across math reasoning, code synthesis, and conversational benchmarks demonstrate that HyperDFlash consistently outperforms both the native MTP baseline and vanilla DFlash adaptation, achieving substantial gains in average accepted draft length and decoding speedup. These results validate HC alignment, gated reduction, and targeted distillation for high-performance speculative decoding. △ Less

Submitted 29 June, 2026; v1 submitted 25 June, 2026; originally announced June 2026.

EgoSAT: A Comprehensive Benchmark of Egocentric Streaming Interaction Understanding

Authors: Yijia Lei, Jinzhao Li, Yichi Zhang, Jiacheng Hua, Yin Li, Miao Liu

Abstract: We introduce EgoSAT, the first comprehensive benchmark for egocentric video reasoning in streaming settings, designed to evaluate the capabilities of modern vision-language models (VLMs). The benchmark targets streaming interaction understanding, where video frames arrive sequentially and models must continuously interpret evolving visual context. EgoSAT unifies several previously distinct tasks w… ▽ More We introduce EgoSAT, the first comprehensive benchmark for egocentric video reasoning in streaming settings, designed to evaluate the capabilities of modern vision-language models (VLMs). The benchmark targets streaming interaction understanding, where video frames arrive sequentially and models must continuously interpret evolving visual context. EgoSAT unifies several previously distinct tasks within a single streaming framework. In this formulation, queries about completed events correspond to retrospective reasoning, queries about ongoing activities require online understanding, and queries about future actions involve prospective anticipation. This unified setting requires models to reason about the past, present, and future while operating under the constraint that only previously observed frames are available. EgoSAT contains 1,997 unique videos spanning 165 hours of egocentric footage and around 4,800 high-quality question-answer pairs, carefully designed to probe reasoning across varying temporal contexts. Using this benchmark, we evaluate a diverse set of both open-weight and closed-weight VLMs, providing a systematic assessment of their ability for streaming interaction understanding. By distinguishing answerability and conducting diagnostics on confidence of models, we find existing models not only struggle with prospective and retrospective modeling, but also exhibit severe mis-calibration: confidence often fails to track inherent answerability, leading to dangerous "confidently wrong" behaviors. Project page: https://leiyj23.github.io/EgoSAT/ △ Less

Submitted 23 June, 2026; originally announced June 2026.

Comments: Accepted to ECCV 2026. Project page: https://leiyj23.github.io/EgoSAT/

EgoIntrospect: An Egocentric Dataset and Benchmark for User-Centric Internal State Reasoning

Authors: Zeyu Wang, Chang Liu, Eduardus Tjitrahardja, Yuntao Wang, Borislav Pavlov, Fangfei Gou, Jose Manuel Davila, Dai Shi, Ran Xu, Yue Pan, Jiayi Tan, Shuting Chang, Qi Wang, Jinzhao Li, Jiacheng Hua, Yifei Huang, Jingwei Sun, Yu Zhang, Liuxin Zhang, Guocai Yao, Jia Jia, Yin Li, Qianying Wang, Yuanchun Shi, Miao Liu

Abstract: Despite extensive efforts on egocentric video datasets and benchmarks, understanding users' internal states, which is crucial for enabling seamless AI assistant experiences, remains largely overlooked. In this work, we introduce EgoIntrospect, the first egocentric dataset captured in user-driven scenarios with self-annotations that explicitly reveal users' interactive intentions with AI assistants… ▽ More Despite extensive efforts on egocentric video datasets and benchmarks, understanding users' internal states, which is crucial for enabling seamless AI assistant experiences, remains largely overlooked. In this work, we introduce EgoIntrospect, the first egocentric dataset captured in user-driven scenarios with self-annotations that explicitly reveal users' interactive intentions with AI assistants. EgoIntrospect was collected using a cross-device setup, providing synchronized video, audio, gaze, motion, and physiological signals. It consists of 180 hours of recordings from 60 subjects, with an average recording duration of 3 hours per subject. Leveraging EgoIntrospect, we formalize a suite of tasks centered on user internal states, including affective experience, interactive intent, and cognitive memory. We further process the annotations to construct benchmarks that evaluate the ability of modern multimodal large language models to reason about users' internal states from egocentric observations. Experiments on our benchmark suggest that existing multimodal large language models struggle to effectively leverage multimodal signals to infer users' subjective internal states. The dataset and annotations will be made publicly available to advance research in egocentric vision and wearable AI assistants. Project page: https://ego-introspect.github.io/ △ Less

Submitted 17 May, 2026; originally announced May 2026.

A Proof-of-Concept Study of Multitask Learning for Cranial Synthetic CT Generation Across Heterogeneous MRI Field Strengths

Authors: Zhuoyao Xin, Yiren Zhang, Christopher Wu, Dong Liu, Chunming Gu, Elena Greco, Erik H. Middlebrooks, Jun Hua, Jia Guo

Abstract: Accurate synthesis of computed tomography (CT) images from magnetic resonance imaging (MRI) is clinically valuable for cranial applications such as attenuation correction, radiotherapy planning, and image-guided interventions. However, heterogeneity across MRI field strengths and acquisition protocols limits the generalizability of existing methods. In this study, we formulate cranial CT synthesis… ▽ More Accurate synthesis of computed tomography (CT) images from magnetic resonance imaging (MRI) is clinically valuable for cranial applications such as attenuation correction, radiotherapy planning, and image-guided interventions. However, heterogeneity across MRI field strengths and acquisition protocols limits the generalizability of existing methods. In this study, we formulate cranial CT synthesis as a modular, structurally coupled problem and propose a deep learning framework to improve robustness across heterogeneous MRI conditions. The model is designed to adapt to variations in field strength and imaging protocols while preserving anatomical consistency. Experiments on multi-site datasets demonstrate improved performance and generalization compared with conventional approaches. The proposed method enables reliable CT synthesis across heterogeneous MRI settings, supporting broader clinical translation. △ Less

Submitted 30 April, 2026; originally announced May 2026.

Comments: Published in Medical Physics (2026). DOI: 10.1002/mp.70429

Journal ref: Medical Physics, 53(5): e70429, 2026

Secret Stealing Attacks on Local LLM Fine-Tuning through Supply-Chain Model Code Backdoors

Authors: Zi Li, Tian Zhou, Wenze Li, Jingyu Hua, Yunlong Mao, Sheng Zhong

Abstract: Local fine-tuning datasets routinely contain sensitive secrets such as API keys, personal identifiers, and financial records. Although ''local offline fine-tuning'' is often viewed as a privacy boundary, we reveal that compromised model code is sufficient to steal them. Current passive pretrained-weight poisoning attacks, while effective for natural language, fundamentally fail to capture such spa… ▽ More Local fine-tuning datasets routinely contain sensitive secrets such as API keys, personal identifiers, and financial records. Although ''local offline fine-tuning'' is often viewed as a privacy boundary, we reveal that compromised model code is sufficient to steal them. Current passive pretrained-weight poisoning attacks, while effective for natural language, fundamentally fail to capture such sparse high-entropy targets due to their reliance on probabilistic semantic prefixes. To bridge this gap, we identify and exploit a practical but overlooked supply-chain vector -- model code camouflaged as standard architectural definitions -- to realize a paradigm shift from passive weight poisoning to active execution hijacking. We introduce a deterministic full-chain memorization mechanism: it locks onto token-level secrets in dynamic computation flows via online tensor-rule matching, and leverages value-gradient decoupling to stealthily inject attack gradients, overcoming gradient drowning to force model memorization. Furthermore, we achieve, for the first time, attacker-verifiable secret stealing through black-box queries that precisely distinguishes true leakage from hallucination. Experiments demonstrate that our method achieves over 98\% Strict ASR without compromising the primary task, and can effectively bypass defense measures including DP-SGD, semantic auditing, and code auditing. △ Less

Submitted 30 April, 2026; originally announced April 2026.

Firstborn Advantage in the Ivory Tower: Mass Science, Expanding Scholarly Families, and the Reshaping of Academic Stratification

Authors: Likun Cao, Jie Hua, James Evans

Abstract: This paper investigates the mechanisms underlying scientific stratification in the era of transition from elite to mass science. Existing scholarship has largely examined scientific stratification through the Matthew effect framework at the individual, institutional, and lineage levels, but this theoretical lens has grown limited in today's academic landscape, where mass, team-based, and lab-cente… ▽ More This paper investigates the mechanisms underlying scientific stratification in the era of transition from elite to mass science. Existing scholarship has largely examined scientific stratification through the Matthew effect framework at the individual, institutional, and lineage levels, but this theoretical lens has grown limited in today's academic landscape, where mass, team-based, and lab-centered research has become the dominant mode of knowledge production. As scientists increasingly share institutional and lineage backgrounds, considerable variation within these units remains unexplained. We propose a new framework that integrates concepts and methodological tools from demography into the social study of science. Drawing on the parallel between biological families and scholarly lineages as fundamental units of reproduction, we adapt the concept of birth order to examine how the sequence of doctoral students within a lineage shapes their career trajectories. Using data on more than one million U.S. doctoral graduates, our analysis shows that, much like in biological families, later students systematically perform worse than earlier ones across multiple dimensions of academic achievement, both short and long term. Examining the underlying mechanisms, we find that later students receive less cognitive stimulation from mature scholars and instead more from peers, and specialize in narrower intellectual domains as senior siblings occupy adjacent territories. These factors constrain their intellectual development as independent scholars. By introducing a demographic framework into the study of science, this paper offers a new perspective on scientific stratification and demonstrates how demographic concepts can be fruitfully extended to analyze broader social and epistemic systems. △ Less

Submitted 28 April, 2026; v1 submitted 24 March, 2026; originally announced April 2026.

Comments: 84 pages; 9 figures

PixelSmile: Toward Fine-Grained Facial Expression Editing

Authors: Jiabin Hua, Hengyuan Xu, Aojie Li, Wei Cheng, Gang Yu, Xingjun Ma, Yu-Gang Jiang

Abstract: Fine-grained facial expression editing has long been limited by intrinsic semantic overlap. To address this, we construct the Flex Facial Expression (FFE) dataset with continuous affective annotations and establish FFE-Bench to evaluate structural confusion, editing accuracy, linear controllability, and the trade-off between expression editing and identity preservation. We propose PixelSmile, a di… ▽ More Fine-grained facial expression editing has long been limited by intrinsic semantic overlap. To address this, we construct the Flex Facial Expression (FFE) dataset with continuous affective annotations and establish FFE-Bench to evaluate structural confusion, editing accuracy, linear controllability, and the trade-off between expression editing and identity preservation. We propose PixelSmile, a diffusion framework that disentangles expression semantics via fully symmetric joint training. PixelSmile combines intensity supervision with contrastive learning to produce stronger and more distinguishable expressions, achieving precise and stable linear expression control through textual latent interpolation. Extensive experiments demonstrate that PixelSmile achieves superior disentanglement and robust identity preservation, confirming its effectiveness for continuous, controllable, and fine-grained expression editing, while naturally supporting smooth expression blending. △ Less

Submitted 26 March, 2026; originally announced March 2026.

Comments: 21 Pages; Project Page: https://ammmob.github.io/PixelSmile/ Code: https://github.com/Ammmob/PixelSmile

Unleashing Spatial Reasoning in Multimodal Large Language Models via Textual Representation Guided Reasoning

Authors: Jiacheng Hua, Yishu Yin, Yuhang Wu, Tai Wang, Yifei Huang, Miao Liu

Abstract: Existing Multimodal Large Language Models (MLLMs) struggle with 3D spatial reasoning, as they fail to construct structured abstractions of the 3D environment depicted in video inputs. To bridge this gap, drawing inspiration from cognitive theories of allocentric spatial reasoning, we investigate how to enable MLLMs to model and reason over text-based spatial representations of video. Specifically,… ▽ More Existing Multimodal Large Language Models (MLLMs) struggle with 3D spatial reasoning, as they fail to construct structured abstractions of the 3D environment depicted in video inputs. To bridge this gap, drawing inspiration from cognitive theories of allocentric spatial reasoning, we investigate how to enable MLLMs to model and reason over text-based spatial representations of video. Specifically, we introduce Textual Representation of Allocentric Context from Egocentric Video (TRACE), a prompting method that induces MLLMs to generate text-based representations of 3D environments as intermediate reasoning traces for more accurate spatial question answering. TRACE encodes meta-context, camera trajectories, and detailed object entities to support structured spatial reasoning over egocentric videos. Extensive experiments on VSI-Bench and OST-Bench demonstrate that TRACE yields notable and consistent improvements over prior prompting strategies across a diverse range of MLLM backbones, spanning different parameter scales and training schemas. We further present ablation studies to validate our design choices, along with detailed analyses that probe the bottlenecks of 3D spatial reasoning in MLLMs. △ Less

Submitted 20 April, 2026; v1 submitted 24 March, 2026; originally announced March 2026.

Comments: Accepted to ACL 2026. 22 pages, 6 figures, 10 tables. Project page: https://trace-reasoning.github.io

Kling-MotionControl Technical Report

Authors: Kling Team, Jialu Chen, Yikang Ding, Zhixue Fang, Kun Gai, Kang He, Xu He, Jingyun Hua, Mingming Lao, Xiaohan Li, Hui Liu, Jiwen Liu, Xiaoqiang Liu, Fan Shi, Xiaoyu Shi, Peiqin Sun, Songlin Tang, Pengfei Wan, Tiancheng Wen, Zhiyong Wu, Haoxian Zhang, Runze Zhao, Yuanxing Zhang, Yan Zhou

Abstract: Character animation aims to generate lifelike videos by transferring motion dynamics from a driving video to a reference image. Recent strides in generative models have paved the way for high-fidelity character animation. In this work, we present Kling-MotionControl, a unified DiT-based framework engineered specifically for robust, precise, and expressive holistic character animation. Leveraging a… ▽ More Character animation aims to generate lifelike videos by transferring motion dynamics from a driving video to a reference image. Recent strides in generative models have paved the way for high-fidelity character animation. In this work, we present Kling-MotionControl, a unified DiT-based framework engineered specifically for robust, precise, and expressive holistic character animation. Leveraging a divide-and-conquer strategy within a cohesive system, the model orchestrates heterogeneous motion representations tailored to the distinct characteristics of body, face, and hands, effectively reconciling large-scale structural stability with fine-grained articulatory expressiveness. To ensure robust cross-identity generalization, we incorporate adaptive identity-agnostic learning, facilitating natural motion retargeting for diverse characters ranging from realistic humans to stylized cartoons. Simultaneously, we guarantee faithful appearance preservation through meticulous identity injection and fusion designs, further supported by a subject library mechanism that leverages comprehensive reference contexts. To ensure practical utility, we implement an advanced acceleration framework utilizing multi-stage distillation, boosting inference speed by over 10x. Kling-MotionControl distinguishes itself through intelligent semantic motion understanding and precise text responsiveness, allowing for flexible control beyond visual inputs. Human preference evaluations demonstrate that Kling-MotionControl delivers superior performance compared to leading commercial and open-source solutions, achieving exceptional fidelity in holistic motion control, open domain generalization, and visual quality and coherence. These results establish Kling-MotionControl as a robust solution for high-quality, controllable, and lifelike character animation. △ Less

Submitted 3 March, 2026; originally announced March 2026.

Comments: Access: https://app.klingai.com/global/video-motion-control/new

Through the Lens of Contrast: Self-Improving Visual Reasoning in VLMs

Authors: Zhiyu Pan, Yizheng Wu, Jiashen Hua, Junyi Feng, Shaotian Yan, Bing Deng, Zhiguo Cao, Jieping Ye

Abstract: Reasoning has emerged as a key capability of large language models. In linguistic tasks, this capability can be enhanced by self-improving techniques that refine reasoning paths for subsequent finetuning. However, extending these language-based self-improving approaches to vision language models (VLMs) presents a unique challenge:~visual hallucinations in reasoning paths cannot be effectively veri… ▽ More Reasoning has emerged as a key capability of large language models. In linguistic tasks, this capability can be enhanced by self-improving techniques that refine reasoning paths for subsequent finetuning. However, extending these language-based self-improving approaches to vision language models (VLMs) presents a unique challenge:~visual hallucinations in reasoning paths cannot be effectively verified or rectified. Our solution starts with a key observation about visual contrast: when presented with a contrastive VQA pair, i.e., two visually similar images with synonymous questions, VLMs identify relevant visual cues more precisely. Motivated by this observation, we propose Visual Contrastive Self-Taught Reasoner (VC-STaR), a novel self-improving framework that leverages visual contrast to mitigate hallucinations in model-generated rationales. We collect a diverse suite of VQA datasets, curate contrastive pairs according to multi-modal similarity, and generate rationales using VC-STaR. Consequently, we obtain a new visual reasoning dataset, VisCoR-55K, which is then used to boost the reasoning capability of various VLMs through supervised finetuning. Extensive experiments show that VC-STaR not only outperforms existing self-improving approaches but also surpasses models finetuned on the SoTA visual reasoning datasets, demonstrating that the inherent contrastive ability of VLMs can bootstrap their own visual reasoning. Project at: https://github.com/zhiyupan42/VC-STaR. △ Less

Submitted 2 March, 2026; originally announced March 2026.

Comments: 19 pages, 9 figures, accepted to ICLR 2026 (oral)

Enhancing Weakly Supervised Multimodal Video Anomaly Detection through Text Guidance

Authors: Shengyang Sun, Jiashen Hua, Junyi Feng, Xiaojin Gong

Abstract: Weakly supervised multimodal video anomaly detection has gained significant attention, yet the potential of the text modality remains under-explored. Text provides explicit semantic information that can enhance anomaly characterization and reduce false alarms. However, extracting effective text features is challenging due to the inability of general-purpose language models to capture anomaly-speci… ▽ More Weakly supervised multimodal video anomaly detection has gained significant attention, yet the potential of the text modality remains under-explored. Text provides explicit semantic information that can enhance anomaly characterization and reduce false alarms. However, extracting effective text features is challenging due to the inability of general-purpose language models to capture anomaly-specific nuances and the scarcity of relevant descriptions. Furthermore, multimodal fusion often suffers from redundancy and imbalance. To address these issues, we propose a novel text-guided framework. First, we introduce an in-context learning-based multi-stage text augmentation mechanism to generate high-quality anomaly text samples for fine-tuning the text feature extractor. Second, we design a multi-scale bottleneck Transformer fusion module that uses compressed bottleneck tokens to progressively integrate information across modalities, mitigating redundancy and imbalance. Experiments on UCF-Crime and XD-Violence demonstrate state-of-the-art performance. △ Less

Submitted 11 February, 2026; originally announced February 2026.

Comments: Accepted by IEEE Transactions on Multimedia

Tackling air quality with SAPIENS

Authors: Marcella Bona, Nathan Heatley, Jia-Chen Hua, Adriana Lara, Valeria Legaria-Santiago, Alberto Luviano Juarez, Fernando Moreno-Gomez, Jocelyn Richardson, Natan Vilchis, Xiwen Shirley Zheng

Abstract: Air pollution is a chronic problem in large cities worldwide and awareness is rising as the long-term health implications become clearer. Vehicular traffic has been identified as a major contributor to poor air quality. In a lot of cities the publicly available air quality measurements and forecasts are coarse-grained both in space and time. However, in general, real-time traffic intensity data is… ▽ More Air pollution is a chronic problem in large cities worldwide and awareness is rising as the long-term health implications become clearer. Vehicular traffic has been identified as a major contributor to poor air quality. In a lot of cities the publicly available air quality measurements and forecasts are coarse-grained both in space and time. However, in general, real-time traffic intensity data is openly available in various forms and is fine-grained. In this paper, we present an in-depth study of pollution sensor measurements combined with traffic data from Mexico City. We analyse and model the relationship between traffic intensity and air quality with the aim to provide hyper-local, dynamic air quality forecasts. We developed an innovative method to represent traffic intensities by transforming simple colour-coded traffic maps into concentric ring-based descriptions, enabling improved characterisation of traffic conditions. Using Partial Least Squares Regression, we predict pollution levels based on these newly defined traffic intensities. The model was optimised with various training samples to achieve the best predictive performance and gain insights into the relationship between pollutants and traffic. The workflow we have designed is straightforward and adaptable to other contexts, like other cities beyond the specifics of our dataset. △ Less

Submitted 30 January, 2026; originally announced January 2026.

Comments: 24 pages, 13 figures

DiaDem: Advancing Dialogue Descriptions in Audiovisual Video Captioning for Multimodal Large Language Models

Authors: Xinlong Chen, Weihong Lin, Jingyun Hua, Linli Yao, Yue Ding, Bozhou Li, Bohan Zeng, Yang Shi, Qiang Liu, Yuanxing Zhang, Pengfei Wan, Liang Wang, Tieniu Tan

Abstract: Accurate dialogue description in audiovisual video captioning is crucial for downstream understanding and generation tasks. However, existing models generally struggle to produce faithful dialogue descriptions within audiovisual captions. To mitigate this limitation, we propose DiaDem, a powerful audiovisual video captioning model capable of generating captions with more precise dialogue descripti… ▽ More Accurate dialogue description in audiovisual video captioning is crucial for downstream understanding and generation tasks. However, existing models generally struggle to produce faithful dialogue descriptions within audiovisual captions. To mitigate this limitation, we propose DiaDem, a powerful audiovisual video captioning model capable of generating captions with more precise dialogue descriptions while maintaining strong overall performance. We first synthesize a high-quality dataset for SFT, then employ a difficulty-partitioned two-stage GRPO strategy to further enhance dialogue descriptions. To enable systematic evaluation of dialogue description capabilities, we introduce DiaDemBench, a comprehensive benchmark designed to evaluate models across diverse dialogue scenarios, emphasizing both speaker attribution accuracy and utterance transcription fidelity in audiovisual captions. Extensive experiments on DiaDemBench reveal even commercial models still exhibit substantial room for improvement in dialogue-aware captioning. Notably, DiaDem not only outperforms the Gemini series in dialogue description accuracy but also achieves competitive performance on general audiovisual captioning benchmarks, demonstrating its overall effectiveness. △ Less

Submitted 27 January, 2026; originally announced January 2026.

Comments: Project webpage: https://diadem-captioner.github.io/

A Safety Report on GPT-5.2, Gemini 3 Pro, Qwen3-VL, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5

Authors: Xingjun Ma, Yixu Wang, Hengyuan Xu, Yutao Wu, Yifan Ding, Yunhan Zhao, Zilong Wang, Jiabin Hua, Ming Wen, Jianan Liu, Ranjie Duan, Yifeng Gao, Yingshui Tan, Yunhao Chen, Hui Xue, Xin Wang, Wei Cheng, Jingjing Chen, Zuxuan Wu, Bo Li, Yu-Gang Jiang

Abstract: The rapid evolution of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has driven major gains in reasoning, perception, and generation across language and vision, yet whether these advances translate into comparable improvements in safety remains unclear, partly due to fragmented evaluations that focus on isolated modalities or threat models. In this report, we present an… ▽ More The rapid evolution of Large Language Models (LLMs) and Multimodal Large Language Models (MLLMs) has driven major gains in reasoning, perception, and generation across language and vision, yet whether these advances translate into comparable improvements in safety remains unclear, partly due to fragmented evaluations that focus on isolated modalities or threat models. In this report, we present an integrated safety evaluation of six frontier models--GPT-5.2, Gemini 3 Pro, Qwen3-VL, Grok 4.1 Fast, Nano Banana Pro, and Seedream 4.5--assessing each across language, vision-language, and image generation using a unified protocol that combines benchmark, adversarial, multilingual, and compliance evaluations. By aggregating results into safety leaderboards and model profiles, we reveal a highly uneven safety landscape: while GPT-5.2 demonstrates consistently strong and balanced performance, other models exhibit clear trade-offs across benchmark safety, adversarial robustness, multilingual generalization, and regulatory compliance. Despite strong results under standard benchmarks, all models remain highly vulnerable under adversarial testing, with worst-case safety rates dropping below 6%. Text-to-image models show slightly stronger alignment in regulated visual risk categories, yet remain fragile when faced with adversarial or semantically ambiguous prompts. Overall, these findings highlight that safety in frontier models is inherently multidimensional--shaped by modality, language, and evaluation design--underscoring the need for standardized, holistic safety assessments to better reflect real-world risk and guide responsible deployment. △ Less

Submitted 16 January, 2026; v1 submitted 15 January, 2026; originally announced January 2026.

Comments: 41 pages, 22 figures

Towards Automated Differential Diagnosis of Skin Diseases Using Deep Learning and Imbalance-Aware Strategies

Authors: Ali Anaissi, Ali Braytee, Weidong Huang, Junaid Akram, Alaa Farhat, Jie Hua

Abstract: As dermatological conditions become increasingly common and the availability of dermatologists remains limited, there is a growing need for intelligent tools to support both patients and clinicians in the timely and accurate diagnosis of skin diseases. In this project, we developed a deep learning based model for the classification and diagnosis of skin conditions. By leveraging pretraining on pub… ▽ More As dermatological conditions become increasingly common and the availability of dermatologists remains limited, there is a growing need for intelligent tools to support both patients and clinicians in the timely and accurate diagnosis of skin diseases. In this project, we developed a deep learning based model for the classification and diagnosis of skin conditions. By leveraging pretraining on publicly available skin disease image datasets, our model effectively extracted visual features and accurately classified various dermatological cases. Throughout the project, we refined the model architecture, optimized data preprocessing workflows, and applied targeted data augmentation techniques to improve overall performance. The final model, based on the Swin Transformer, achieved a prediction accuracy of 87.71 percent across eight skin lesion classes on the ISIC2019 dataset. These results demonstrate the model's potential as a diagnostic support tool for clinicians and a self assessment aid for patients. △ Less

Submitted 1 January, 2026; originally announced January 2026.

Comments: The 23rd Australasian Data Science and Machine Learning Conference (AusDM'25)

Benchmarking Preprocessing and Integration Methods in Single-Cell Genomics

Authors: Ali Anaissi, Seid Miad Zandavi, Weidong Huang, Junaid Akram, Basem Suleiman, Ali Braytee, Jie Hua

Abstract: Single-cell data analysis has the potential to revolutionize personalized medicine by characterizing disease-associated molecular changes at the single-cell level. Advanced single-cell multimodal assays can now simultaneously measure various molecules (e.g., DNA, RNA, Protein) across hundreds of thousands of individual cells, providing a comprehensive molecular readout. A significant analytical ch… ▽ More Single-cell data analysis has the potential to revolutionize personalized medicine by characterizing disease-associated molecular changes at the single-cell level. Advanced single-cell multimodal assays can now simultaneously measure various molecules (e.g., DNA, RNA, Protein) across hundreds of thousands of individual cells, providing a comprehensive molecular readout. A significant analytical challenge is integrating single-cell measurements across different modalities. Various methods have been developed to address this challenge, but there has been no systematic evaluation of these techniques with different preprocessing strategies. This study examines a general pipeline for single-cell data analysis, which includes normalization, data integration, and dimensionality reduction. The performance of different algorithm combinations often depends on the dataset sizes and characteristics. We evaluate six datasets across diverse modalities, tissues, and organisms using three metrics: Silhouette Coefficient Score, Adjusted Rand Index, and Calinski-Harabasz Index. Our experiments involve combinations of seven normalization methods, four dimensional reduction methods, and five integration methods. The results show that Seurat and Harmony excel in data integration, with Harmony being more time-efficient, especially for large datasets. UMAP is the most compatible dimensionality reduction method with the integration techniques, and the choice of normalization method varies depending on the integration method used. △ Less

Submitted 1 January, 2026; originally announced January 2026.

Comments: The 23rd Australasian Data Science and Machine Learning Conference (AusDM'25)

KlingAvatar 2.0 Technical Report

Authors: Kling Team, Jialu Chen, Yikang Ding, Zhixue Fang, Kun Gai, Yuan Gao, Kang He, Jingyun Hua, Boyuan Jiang, Mingming Lao, Xiaohan Li, Hui Liu, Jiwen Liu, Xiaoqiang Liu, Yuan Liu, Shun Lu, Yongsen Mao, Yingchao Shao, Huafeng Shi, Xiaoyu Shi, Peiqin Sun, Songlin Tang, Pengfei Wan, Chao Wang, Xuebo Wang , et al. (3 additional authors not shown)

Abstract: Avatar video generation models have achieved remarkable progress in recent years. However, prior work exhibits limited efficiency in generating long-duration high-resolution videos, suffering from temporal drifting, quality degradation, and weak prompt following as video length increases. To address these challenges, we propose KlingAvatar 2.0, a spatio-temporal cascade framework that performs ups… ▽ More Avatar video generation models have achieved remarkable progress in recent years. However, prior work exhibits limited efficiency in generating long-duration high-resolution videos, suffering from temporal drifting, quality degradation, and weak prompt following as video length increases. To address these challenges, we propose KlingAvatar 2.0, a spatio-temporal cascade framework that performs upscaling in both spatial resolution and temporal dimension. The framework first generates low-resolution blueprint video keyframes that capture global semantics and motion, and then refines them into high-resolution, temporally coherent sub-clips using a first-last frame strategy, while retaining smooth temporal transitions in long-form videos. To enhance cross-modal instruction fusion and alignment in extended videos, we introduce a Co-Reasoning Director composed of three modality-specific large language model (LLM) experts. These experts reason about modality priorities and infer underlying user intent, converting inputs into detailed storylines through multi-turn dialogue. A Negative Director further refines negative prompts to improve instruction alignment. Building on these components, we extend the framework to support ID-specific multi-character control. Extensive experiments demonstrate that our model effectively addresses the challenges of efficient, multimodally aligned long-form high-resolution video generation, delivering enhanced visual clarity, realistic lip-teeth rendering with accurate lip synchronization, strong identity preservation, and coherent multimodal instruction following. △ Less

Submitted 15 December, 2025; originally announced December 2025.

Comments: 14 pages, 7 figures

Distillability of LLM Security Logic: Predicting Attack Success Rate of Outline Filling Attack via Ranking Regression

Authors: Tianyu Zhang, Zihang Xi, Jingyu Hua, Sheng Zhong

Abstract: In the realm of black-box jailbreak attacks on large language models (LLMs), the feasibility of constructing a narrow safety proxy, a lightweight model designed to predict the attack success rate (ASR) of adversarial prompts, remains underexplored. This work investigates the distillability of an LLM's core security logic. We propose a novel framework that incorporates an improved outline filling a… ▽ More In the realm of black-box jailbreak attacks on large language models (LLMs), the feasibility of constructing a narrow safety proxy, a lightweight model designed to predict the attack success rate (ASR) of adversarial prompts, remains underexplored. This work investigates the distillability of an LLM's core security logic. We propose a novel framework that incorporates an improved outline filling attack to achieve dense sampling of the model's security boundaries. Furthermore, we introduce a ranking regression paradigm that replaces standard regression and trains the proxy model to predict which prompt yields a higher ASR. Experimental results show that our proxy model achieves an accuracy of 91.1 percent in predicting the relative ranking of average long response (ALR), and 69.2 percent in predicting ASR. These findings confirm the predictability and distillability of jailbreak behaviors, and demonstrate the potential of leveraging such distillability to optimize black-box attacks. △ Less

Submitted 26 November, 2025; originally announced November 2025.

Mind the Gap: Revealing Inconsistencies Across Heterogeneous AI Accelerators

Authors: Elliott Wen, Sean Ma, Ewan Tempero, Jens Dietrich, Daniel Luo, Jiaxing Shen, Kaiqi Zhao, Bruce Sham, Yousong Song, Jiayi Hua, Jia Hong

Abstract: While NVIDIA remains the dominant provider of AI accelerators within cloud data center, emerging vendors such as AMD, Intel, Mac, and Huawei offer cost-effective alternatives with claims of compatibility and performance. This paper presents the first empirical study investigating divergence in machine learning model across heterogeneous AI accelerators. Utilizing an automated pipeline, we synthesi… ▽ More While NVIDIA remains the dominant provider of AI accelerators within cloud data center, emerging vendors such as AMD, Intel, Mac, and Huawei offer cost-effective alternatives with claims of compatibility and performance. This paper presents the first empirical study investigating divergence in machine learning model across heterogeneous AI accelerators. Utilizing an automated pipeline, we synthesize over 100,000 variant models derived from 4,000 real-world models and execute them across five different enterprise-grade accelerators. Our findings suggest that newer AI platforms from Mac and Huawei support at least 17\% fewer operators than NVIDIA. These platforms also exhibit a higher rate of output discrepancies (exceeding 5\%), which stem from differences in operator implementations, handling of exceptional numerical values, and instruction scheduling. They are also more susceptible to failures during model compilation-based acceleration, and in some cases, the compiled models produce outputs that differ noticeably from those generated using the standard execution mode. In addition, we identify 7 implementation flaws in PyTorch and 40 platform-specific issues across vendors. These results underscore the challenges of achieving consistent machine learning behavior in an increasingly diverse hardware ecosystem. △ Less

Submitted 30 October, 2025; originally announced November 2025.

CoDefend: Cross-Modal Collaborative Defense via Diffusion Purification and Prompt Optimization

Authors: Fengling Zhu, Boshi Liu, Jingyu Hua, Sheng Zhong

Abstract: Multimodal Large Language Models (MLLMs) have achieved remarkable success in tasks such as image captioning, visual question answering, and cross-modal reasoning by integrating visual and textual modalities. However, their multimodal nature also exposes them to adversarial threats, where attackers can perturb either modality or both jointly to induce harmful, misleading, or policy violating output… ▽ More Multimodal Large Language Models (MLLMs) have achieved remarkable success in tasks such as image captioning, visual question answering, and cross-modal reasoning by integrating visual and textual modalities. However, their multimodal nature also exposes them to adversarial threats, where attackers can perturb either modality or both jointly to induce harmful, misleading, or policy violating outputs. Existing defense strategies, such as adversarial training and input purification, face notable limitations: adversarial training typically improves robustness only against known attacks while incurring high computational costs, whereas conventional purification approaches often suffer from degraded image quality and insufficient generalization to complex multimodal tasks. In this work, we focus on defending the visual modality, which frequently serves as the primary entry point for adversarial manipulation. We propose a supervised diffusion based denoising framework that leverages paired adversarial clean image datasets to fine-tune diffusion models with directional, task specific guidance. Unlike prior unsupervised purification methods such as DiffPure, our approach achieves higher quality reconstructions while significantly improving defense robustness in multimodal tasks. Furthermore, we incorporate prompt optimization as a complementary defense mechanism, enhancing resistance against diverse and unseen attack strategies. Extensive experiments on image captioning and visual question answering demonstrate that our method not only substantially improves robustness but also exhibits strong transferability to unknown adversarial attacks. These results highlight the effectiveness of supervised diffusion based denoising for multimodal defense, paving the way for more reliable and secure deployment of MLLMs in real world applications. △ Less

Submitted 13 October, 2025; originally announced October 2025.

AVoCaDO: An Audiovisual Video Captioner Driven by Temporal Orchestration

Authors: Xinlong Chen, Yue Ding, Weihong Lin, Jingyun Hua, Linli Yao, Yang Shi, Bozhou Li, Yuanxing Zhang, Qiang Liu, Pengfei Wan, Liang Wang, Tieniu Tan

Abstract: Audiovisual video captioning aims to generate semantically rich descriptions with temporal alignment between visual and auditory events, thereby benefiting both video understanding and generation. In this paper, we present AVoCaDO, a powerful audiovisual video captioner driven by the temporal orchestration between audio and visual modalities. We propose a two-stage post-training pipeline: (1) AVoC… ▽ More Audiovisual video captioning aims to generate semantically rich descriptions with temporal alignment between visual and auditory events, thereby benefiting both video understanding and generation. In this paper, we present AVoCaDO, a powerful audiovisual video captioner driven by the temporal orchestration between audio and visual modalities. We propose a two-stage post-training pipeline: (1) AVoCaDO SFT, which fine-tunes the model on a newly curated dataset of 107K high-quality, temporally-aligned audiovisual captions; and (2) AVoCaDO GRPO, which leverages tailored reward functions to further enhance temporal coherence and dialogue accuracy while regularizing caption length and reducing collapse. Experimental results demonstrate that AVoCaDO significantly outperforms existing open-source models across four audiovisual video captioning benchmarks, and also achieves competitive performance on the VDC and DREAM-1K benchmark under visual-only settings. △ Less

Submitted 11 October, 2025; originally announced October 2025.

Comments: Project webpage: https://avocado-captioner.github.io/

Extreme value forecasting using relevance-based data augmentation with deep learning models

Authors: Junru Hua, Rahul Ahluwalia, Rohitash Chandra

Abstract: Data augmentation with generative adversarial networks (GANs) has been popular for class imbalance problems, mainly for pattern classification and computer vision-related applications. Extreme value forecasting is a challenging field that has various applications from finance to climate change problems. In this study, we present a data augmentation framework for extreme value forecasting. In this… ▽ More Data augmentation with generative adversarial networks (GANs) has been popular for class imbalance problems, mainly for pattern classification and computer vision-related applications. Extreme value forecasting is a challenging field that has various applications from finance to climate change problems. In this study, we present a data augmentation framework for extreme value forecasting. In this framework, our focus is on forecasting extreme values using deep learning models in combination with data augmentation models such as GANs and synthetic minority oversampling technique (SMOTE). We use deep learning models such as convolutional long short-term memory (Conv-LSTM) and bidirectional long short-term memory (BD-LSTM) networks for multistep ahead prediction featuring extremes. We investigate which data augmentation models are the most suitable, taking into account the prediction accuracy overall and at extreme regions, along with computational efficiency. We also present novel strategies for incorporating data augmentation, considering extreme values based on a relevance function. Our results indicate that the SMOTE-based strategy consistently demonstrated superior adaptability, leading to improved performance across both short- and long-horizon forecasts. Conv-LSTM and BD-LSTM exhibit complementary strengths: the former excels in periodic, stable datasets, while the latter performs better in chaotic or non-stationary sequences. △ Less

Submitted 2 October, 2025; originally announced October 2025.

Synergistic Enhancement of Requirement-to-Code Traceability: A Framework Combining Large Language Model based Data Augmentation and an Advanced Encoder

Authors: Jianzhang Zhang, Jialong Zhou, Nan Niu, Jinping Hua, Chuang Liu

Abstract: Automated requirement-to-code traceability link recovery, essential for industrial system quality and safety, is critically hindered by the scarcity of labeled data. To address this bottleneck, this paper proposes and validates a synergistic framework that integrates large language model (LLM)-driven data augmentation with an advanced encoder. We first demonstrate that data augmentation, optimized… ▽ More Automated requirement-to-code traceability link recovery, essential for industrial system quality and safety, is critically hindered by the scarcity of labeled data. To address this bottleneck, this paper proposes and validates a synergistic framework that integrates large language model (LLM)-driven data augmentation with an advanced encoder. We first demonstrate that data augmentation, optimized through a systematic evaluation of bi-directional and zero/few-shot prompting strategies, is highly effective, while the choice among leading LLMs is not a significant performance factor. Building on the augmented data, we further enhance an established, state-of-the-art pre-trained language model based method by incorporating an encoder distinguished by a broader pre-training corpus and an extended context window. Our experiments on four public datasets quantify the distinct contributions of our framework's components: on its own, data augmentation consistently improves the baseline method, providing substantial performance gains of up to 26.66%; incorporating the advanced encoder provides an additional lift of 2.21% to 11.25%. This synergy culminates in a fully optimized framework with maximum gains of up to 28.59% on $F_1$ score and 28.9% on $F_2$ score over the established baseline, decisively outperforming ten established baselines from three dominant paradigms. This work contributes a pragmatic and scalable methodology to overcome the data scarcity bottleneck, paving the way for broader industrial adoption of data-driven requirement-to-code traceability. △ Less

Submitted 19 October, 2025; v1 submitted 24 September, 2025; originally announced September 2025.

In the Eye of MLLM: Benchmarking Egocentric Video Intent Understanding with Gaze-Guided Prompting

Authors: Taiying Peng, Jiacheng Hua, Miao Liu, Feng Lu

Abstract: The emergence of advanced multimodal large language models (MLLMs) has significantly enhanced AI assistants' ability to process complex information across modalities. Recently, egocentric videos, by directly capturing user focus, actions, and context in an unified coordinate, offer an exciting opportunity to enable proactive and personalized AI user experiences with MLLMs. However, existing benchm… ▽ More The emergence of advanced multimodal large language models (MLLMs) has significantly enhanced AI assistants' ability to process complex information across modalities. Recently, egocentric videos, by directly capturing user focus, actions, and context in an unified coordinate, offer an exciting opportunity to enable proactive and personalized AI user experiences with MLLMs. However, existing benchmarks overlook the crucial role of gaze as an indicator of user intent. To address this gap, we introduce EgoGazeVQA, an egocentric gaze-guided video question answering benchmark that leverages gaze information to improve the understanding of longer daily-life videos. EgoGazeVQA consists of gaze-based QA pairs generated by MLLMs and refined by human annotators. Our experiments reveal that existing MLLMs struggle to accurately interpret user intentions. In contrast, our gaze-guided intent prompting methods significantly enhance performance by integrating spatial, temporal, and intent-related cues. We further conduct experiments on gaze-related fine-tuning and analyze how gaze estimation accuracy impacts prompting effectiveness. These results underscore the value of gaze for more personalized and effective AI assistants in egocentric settings. Project page: https://taiyi98.github.io/projects/EgoGazeVQA △ Less

Submitted 14 October, 2025; v1 submitted 9 September, 2025; originally announced September 2025.

Comments: Accepted to NeurIPS 2025

Kwai Keye-VL 1.5 Technical Report

Authors: Biao Yang, Bin Wen, Boyang Ding, Changyi Liu, Chenglong Chu, Chengru Song, Chongling Rao, Chuan Yi, Da Li, Dunju Zang, Fan Yang, Guorui Zhou, Guowang Zhang, Han Shen, Hao Peng, Haojie Ding, Hao Wang, Haonan Fan, Hengrui Ju, Jiaming Huang, Jiangxia Cao, Jiankang Chen, Jingyun Hua, Kaibing Chen, Kaiyu Jiang , et al. (36 additional authors not shown)

Abstract: In recent years, the development of Large Language Models (LLMs) has significantly advanced, extending their capabilities to multimodal tasks through Multimodal Large Language Models (MLLMs). However, video understanding remains a challenging area due to the dynamic and information-dense nature of videos. Existing models struggle with the trade-off between spatial resolution and temporal coverage… ▽ More In recent years, the development of Large Language Models (LLMs) has significantly advanced, extending their capabilities to multimodal tasks through Multimodal Large Language Models (MLLMs). However, video understanding remains a challenging area due to the dynamic and information-dense nature of videos. Existing models struggle with the trade-off between spatial resolution and temporal coverage when processing video content. We present Keye-VL-1.5, which addresses fundamental challenges in video comprehension through three key innovations. First, we introduce a novel Slow-Fast video encoding strategy that dynamically allocates computational resources based on inter-frame similarity, processing key frames with significant visual changes at higher resolution (Slow pathway) while handling relatively static frames with increased temporal coverage at lower resolution (Fast pathway). Second, we implement a progressive four-stage pre-training methodology that systematically extends the model's context length from 8K to 128K tokens, enabling processing of longer videos and more complex visual content. Third, we develop a comprehensive post-training pipeline focusing on reasoning enhancement and human preference alignment, incorporating a 5-step chain-of-thought data construction process, iterative GSPO-based reinforcement learning with progressive prompt hinting for difficult cases, and alignment training. Through extensive evaluation on public benchmarks and rigorous internal human assessment, Keye-VL-1.5 demonstrates significant improvements over existing models, particularly excelling in video understanding tasks while maintaining competitive performance on general multimodal benchmarks. △ Less

Submitted 7 September, 2025; v1 submitted 1 September, 2025; originally announced September 2025.

Comments: Github page: https://github.com/Kwai-Keye/Keye

CS-Agent: LLM-based Community Search via Dual-agent Collaboration

Authors: Jiahao Hua, Long Yuan, Qingshuai Feng, Qiang Fan, Shan Huang

Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, yet their application to graph structure analysis, particularly in community search, remains underexplored. Community search, a fundamental task in graph analysis, aims to identify groups of nodes with dense interconnections, which is crucial for understanding the macroscopic structure of g… ▽ More Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing tasks, yet their application to graph structure analysis, particularly in community search, remains underexplored. Community search, a fundamental task in graph analysis, aims to identify groups of nodes with dense interconnections, which is crucial for understanding the macroscopic structure of graphs. In this paper, we propose GraphCS, a comprehensive benchmark designed to evaluate the performance of LLMs in community search tasks. Our experiments reveal that while LLMs exhibit preliminary potential, they frequently fail to return meaningful results and suffer from output bias. To address these limitations, we introduce CS-Agent, a dual-agent collaborative framework to enhance LLM-based community search. CS-Agent leverages the complementary strengths of two LLMs acting as Solver and Validator. Through iterative feedback and refinement, CS-Agent dynamically refines initial results without fine-tuning or additional training. After the multi-round dialogue, Decider module selects the optimal community. Extensive experiments demonstrate that CS-Agent significantly improves the quality and stability of identified communities compared to baseline methods. To our knowledge, this is the first work to apply LLMs to community search, bridging the gap between LLMs and graph analysis while providing a robust and adaptive solution for real-world applications. △ Less

Submitted 14 August, 2025; v1 submitted 13 August, 2025; originally announced August 2025.

Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment

Authors: Dai Li, Kevin Course, Wei Li, Hongwei Li, Jie Hua, Yiqi Chen, Zhao Zhu, Rui Jian, Xuan Cao, Bi Xue, Yu Shi, Jing Qian, Kai Ren, Matt Ma, Qunshu Zhang, Rui Li

Abstract: Scaling laws have been established for recommender systems, yet efficiently deploying foundation model (FM) across multiple recommendation surfaces remains a major unsolved challenge. Existing methods for transfer learning face fundamental limitations in this setting: knowledge distillation suffers from transfer fidelity degradation in the large-data regime, and static user or item embeddings lack… ▽ More Scaling laws have been established for recommender systems, yet efficiently deploying foundation model (FM) across multiple recommendation surfaces remains a major unsolved challenge. Existing methods for transfer learning face fundamental limitations in this setting: knowledge distillation suffers from transfer fidelity degradation in the large-data regime, and static user or item embeddings lack the expressiveness to capture contextualized user-item interactions. We propose the Foundation-Expert paradigm, where a central FM generates target-aware embeddings which are ingested by lightweight surface-specific expert models. Target-aware embeddings are representations that dynamically capture a user's interest in a specific item conditioned on their full interaction history. Unlike knowledge distillation, which transfers FM knowledge as soft labels, our approach provides these embeddings as input features to each expert model, enabling direct interaction with surface-specific representations. This paradigm achieves transfer ratios of 0.64--1.0 from FM to experts, substantially exceeding existing methods. Fully deployed at Meta serving tens of billions of daily requests since 2025, it delivers 0.050% statistically significant online topline metric improvement and 0.359% cumulative gains across multiple surfaces. △ Less

Submitted 24 July, 2026; v1 submitted 4 August, 2025; originally announced August 2025.

MSC Class: 68T05; 68T07; 68T30 ACM Class: H.3.3; I.2.6

Kwai Keye-VL Technical Report

Authors: Kwai Keye Team, Biao Yang, Bin Wen, Changyi Liu, Chenglong Chu, Chengru Song, Chongling Rao, Chuan Yi, Da Li, Dunju Zang, Fan Yang, Guorui Zhou, Hao Peng, Haojie Ding, Jiaming Huang, Jiangxia Cao, Jiankang Chen, Jingyun Hua, Jin Ouyang, Kaibing Chen, Kaiyu Jiang, Kaiyu Tang, Kun Gai, Shengnan Zhang, Siyang Mao , et al. (35 additional authors not shown)

Abstract: While Multimodal Large Language Models (MLLMs) demonstrate remarkable capabilities on static images, they often fall short in comprehending dynamic, information-dense short-form videos, a dominant medium in today's digital landscape. To bridge this gap, we introduce \textbf{Kwai Keye-VL}, an 8-billion-parameter multimodal foundation model engineered for leading-edge performance in short-video unde… ▽ More While Multimodal Large Language Models (MLLMs) demonstrate remarkable capabilities on static images, they often fall short in comprehending dynamic, information-dense short-form videos, a dominant medium in today's digital landscape. To bridge this gap, we introduce \textbf{Kwai Keye-VL}, an 8-billion-parameter multimodal foundation model engineered for leading-edge performance in short-video understanding while maintaining robust general-purpose vision-language abilities. The development of Keye-VL rests on two core pillars: a massive, high-quality dataset exceeding 600 billion tokens with a strong emphasis on video, and an innovative training recipe. This recipe features a four-stage pre-training process for solid vision-language alignment, followed by a meticulous two-phase post-training process. The first post-training stage enhances foundational capabilities like instruction following, while the second phase focuses on stimulating advanced reasoning. In this second phase, a key innovation is our five-mode ``cold-start'' data mixture, which includes ``thinking'', ``non-thinking'', ``auto-think'', ``think with image'', and high-quality video data. This mixture teaches the model to decide when and how to reason. Subsequent reinforcement learning (RL) and alignment steps further enhance these reasoning capabilities and correct abnormal model behaviors, such as repetitive outputs. To validate our approach, we conduct extensive evaluations, showing that Keye-VL achieves state-of-the-art results on public video benchmarks and remains highly competitive on general image-based tasks (Figure 1). Furthermore, we develop and release the \textbf{KC-MMBench}, a new benchmark tailored for real-world short-video scenarios, where Keye-VL shows a significant advantage. △ Less

Submitted 2 July, 2025; originally announced July 2025.

Comments: Technical Report: https://github.com/Kwai-Keye/Keye

Designing 3D Anisotropic Frame Fields with Odeco Tensors

Authors: Haikuan Zhu, Hongbo Li, Hsueh-Ti Derek Liu, Wenping Wang, Jing Hua, Zichun Zhong

Abstract: This paper introduces a method to synthesize a 3D tensor field within a constrained geometric domain represented as a tetrahedral mesh. Whereas previous techniques optimize for isotropic fields, we focus on anisotropic tensor fields that are smooth and aligned with the domain boundary or user guidance. The key ingredient of our method is a novel computational design framework, built on top of the… ▽ More This paper introduces a method to synthesize a 3D tensor field within a constrained geometric domain represented as a tetrahedral mesh. Whereas previous techniques optimize for isotropic fields, we focus on anisotropic tensor fields that are smooth and aligned with the domain boundary or user guidance. The key ingredient of our method is a novel computational design framework, built on top of the symmetric orthogonally decomposable (odeco) tensor representation, to optimize the stretching ratios and orientations for each tensor in the domain. In contrast to past techniques designed only for isotropic tensors, we demonstrate the efficacy of our approach in generating smooth volumetric tensor fields with high anisotropy and shape conformity, especially for the domain with complex shapes. We apply these anisotropic tensor fields to various applications, such as anisotropic meshing, structural mechanics, and fabrication. △ Less

Submitted 8 May, 2025; originally announced May 2025.

Comments: it was accepted by TOG

Mavors: Multi-granularity Video Representation for Multimodal Large Language Model

Authors: Yang Shi, Jiaheng Liu, Yushuo Guan, Zhenhua Wu, Yuanxing Zhang, Zihao Wang, Weihong Lin, Jingyun Hua, Zekun Wang, Xinlong Chen, Bohan Zeng, Wentao Zhang, Fuzheng Zhang, Wenjing Yang, Di Zhang

Abstract: Long-context video understanding in multimodal large language models (MLLMs) faces a critical challenge: balancing computational efficiency with the retention of fine-grained spatio-temporal patterns. Existing approaches (e.g., sparse sampling, dense sampling with low resolution, and token compression) suffer from significant information loss in temporal dynamics, spatial details, or subtle intera… ▽ More Long-context video understanding in multimodal large language models (MLLMs) faces a critical challenge: balancing computational efficiency with the retention of fine-grained spatio-temporal patterns. Existing approaches (e.g., sparse sampling, dense sampling with low resolution, and token compression) suffer from significant information loss in temporal dynamics, spatial details, or subtle interactions, particularly in videos with complex motion or varying resolutions. To address this, we propose $\mathbf{Mavors}$, a novel framework that introduces $\mathbf{M}$ulti-gr$\mathbf{a}$nularity $\mathbf{v}$ide$\mathbf{o}$ $\mathbf{r}$epre$\mathbf{s}$entation for holistic long-video modeling. Specifically, Mavors directly encodes raw video content into latent representations through two core components: 1) an Intra-chunk Vision Encoder (IVE) that preserves high-resolution spatial features via 3D convolutions and Vision Transformers, and 2) an Inter-chunk Feature Aggregator (IFA) that establishes temporal coherence across chunks using transformer-based dependency modeling with chunk-level rotary position encodings. Moreover, the framework unifies image and video understanding by treating images as single-frame videos via sub-image decomposition. Experiments across diverse benchmarks demonstrate Mavors' superiority in maintaining both spatial fidelity and temporal continuity, significantly outperforming existing methods in tasks requiring fine-grained spatio-temporal reasoning. △ Less

Submitted 26 November, 2025; v1 submitted 14 April, 2025; originally announced April 2025.

Comments: 22 pages

OFL: Opportunistic Federated Learning for Resource-Heterogeneous and Privacy-Aware Devices

Authors: Yunlong Mao, Mingyang Niu, Ziqin Dang, Chengxi Li, Hanning Xia, Yuejuan Zhu, Haoyu Bian, Yuan Zhang, Jingyu Hua, Sheng Zhong

Abstract: Efficient and secure federated learning (FL) is a critical challenge for resource-limited devices, especially mobile devices. Existing secure FL solutions commonly incur significant overhead, leading to a contradiction between efficiency and security. As a result, these two concerns are typically addressed separately. This paper proposes Opportunistic Federated Learning (OFL), a novel FL framework… ▽ More Efficient and secure federated learning (FL) is a critical challenge for resource-limited devices, especially mobile devices. Existing secure FL solutions commonly incur significant overhead, leading to a contradiction between efficiency and security. As a result, these two concerns are typically addressed separately. This paper proposes Opportunistic Federated Learning (OFL), a novel FL framework designed explicitly for resource-heterogenous and privacy-aware FL devices, solving efficiency and security problems jointly. OFL optimizes resource utilization and adaptability across diverse devices by adopting a novel hierarchical and asynchronous aggregation strategy. OFL provides strong security by introducing a differentially private and opportunistic model updating mechanism for intra-cluster model aggregation and an advanced threshold homomorphic encryption scheme for inter-cluster aggregation. Moreover, OFL secures global model aggregation by implementing poisoning attack detection using frequency analysis while keeping models encrypted. We have implemented OFL in a real-world testbed and evaluated OFL comprehensively. The evaluation results demonstrate that OFL achieves satisfying model performance and improves efficiency and security, outperforming existing solutions. △ Less

Submitted 19 March, 2025; originally announced March 2025.

Comments: 14 pages, 13 figures

PSF-4D: A Progressive Sampling Framework for View Consistent 4D Editing

Authors: Hasan Iqbal, Nazmul Karim, Umar Khalid, Azib Farooq, Zichun Zhong, Chen Chen, Jing Hua

Abstract: Instruction-guided generative models, especially those using text-to-image (T2I) and text-to-video (T2V) diffusion frameworks, have advanced the field of content editing in recent years. To extend these capabilities to 4D scene, we introduce a progressive sampling framework for 4D editing (PSF-4D) that ensures temporal and multi-view consistency by intuitively controlling the noise initialization… ▽ More Instruction-guided generative models, especially those using text-to-image (T2I) and text-to-video (T2V) diffusion frameworks, have advanced the field of content editing in recent years. To extend these capabilities to 4D scene, we introduce a progressive sampling framework for 4D editing (PSF-4D) that ensures temporal and multi-view consistency by intuitively controlling the noise initialization during forward diffusion. For temporal coherence, we design a correlated Gaussian noise structure that links frames over time, allowing each frame to depend meaningfully on prior frames. Additionally, to ensure spatial consistency across views, we implement a cross-view noise model, which uses shared and independent noise components to balance commonalities and distinct details among different views. To further enhance spatial coherence, PSF-4D incorporates view-consistent iterative refinement, embedding view-aware information into the denoising process to ensure aligned edits across frames and views. Our approach enables high-quality 4D editing without relying on external models, addressing key challenges in previous methods. Through extensive evaluation on multiple benchmarks and multiple editing aspects (e.g., style transfer, multi-attribute editing, object removal, local editing, etc.), we show the effectiveness of our proposed method. Experimental results demonstrate that our proposed method outperforms state-of-the-art 4D editing methods in diverse benchmarks. △ Less

Submitted 31 March, 2025; v1 submitted 13 March, 2025; originally announced March 2025.

Comments: 9 pages, 7 figures

HAIC: Improving Human Action Understanding and Generation with Better Captions for Multi-modal Large Language Models

Authors: Xiao Wang, Jingyun Hua, Weihong Lin, Yuanxing Zhang, Fuzheng Zhang, Jianlong Wu, Di Zhang, Liqiang Nie

Abstract: Recent Multi-modal Large Language Models (MLLMs) have made great progress in video understanding. However, their performance on videos involving human actions is still limited by the lack of high-quality data. To address this, we introduce a two-stage data annotation pipeline. First, we design strategies to accumulate videos featuring clear human actions from the Internet. Second, videos are annot… ▽ More Recent Multi-modal Large Language Models (MLLMs) have made great progress in video understanding. However, their performance on videos involving human actions is still limited by the lack of high-quality data. To address this, we introduce a two-stage data annotation pipeline. First, we design strategies to accumulate videos featuring clear human actions from the Internet. Second, videos are annotated in a standardized caption format that uses human attributes to distinguish individuals and chronologically details their actions and interactions. Through this pipeline, we curate two datasets, namely HAICTrain and HAICBench. \textbf{HAICTrain} comprises 126K video-caption pairs generated by Gemini-Pro and verified for training purposes. Meanwhile, \textbf{HAICBench} includes 412 manually annotated video-caption pairs and 2,000 QA pairs, for a comprehensive evaluation of human action understanding. Experimental results demonstrate that training with HAICTrain not only significantly enhances human understanding abilities across 4 benchmarks, but can also improve text-to-video generation results. Both the HAICTrain and HAICBench are released at https://huggingface.co/datasets/KuaishouHAIC/HAIC. △ Less

Submitted 8 June, 2025; v1 submitted 28 February, 2025; originally announced February 2025.

Self-Instruct Few-Shot Jailbreaking: Decompose the Attack into Pattern and Behavior Learning

Authors: Jiaqi Hua, Wanxu Wei

Abstract: Recently, several works have been conducted on jailbreaking Large Language Models (LLMs) with few-shot malicious demos. In particular, Zheng et al. focus on improving the efficiency of Few-Shot Jailbreaking (FSJ) by injecting special tokens into the demos and employing demo-level random search, known as Improved Few-Shot Jailbreaking (I-FSJ). Nevertheless, we notice that this method may still requ… ▽ More Recently, several works have been conducted on jailbreaking Large Language Models (LLMs) with few-shot malicious demos. In particular, Zheng et al. focus on improving the efficiency of Few-Shot Jailbreaking (FSJ) by injecting special tokens into the demos and employing demo-level random search, known as Improved Few-Shot Jailbreaking (I-FSJ). Nevertheless, we notice that this method may still require a long context to jailbreak advanced models e.g. 32 shots of demos for Meta-Llama-3-8B-Instruct (Llama-3) \cite{llama3modelcard}. In this paper, we discuss the limitations of I-FSJ and propose Self-Instruct Few-Shot Jailbreaking (Self-Instruct-FSJ) facilitated with the demo-level greedy search. This framework decomposes the FSJ attack into pattern and behavior learning to exploit the model's vulnerabilities in a more generalized and efficient way. We conduct elaborate experiments to evaluate our method on common open-source models and compare it with baseline algorithms. Our code is available at https://github.com/iphosi/Self-Instruct-FSJ. △ Less

Submitted 1 February, 2025; v1 submitted 14 January, 2025; originally announced January 2025.

Query Quantized Neural SLAM

Authors: Sijia Jiang, Jing Hua, Zhizhong Han

Abstract: Neural implicit representations have shown remarkable abilities in jointly modeling geometry, color, and camera poses in simultaneous localization and mapping (SLAM). Current methods use coordinates, positional encodings, or other geometry features as input to query neural implicit functions for signed distances and color which produce rendering errors to drive the optimization in overfitting imag… ▽ More Neural implicit representations have shown remarkable abilities in jointly modeling geometry, color, and camera poses in simultaneous localization and mapping (SLAM). Current methods use coordinates, positional encodings, or other geometry features as input to query neural implicit functions for signed distances and color which produce rendering errors to drive the optimization in overfitting image observations. However, due to the run time efficiency requirement in SLAM systems, we are merely allowed to conduct optimization on each frame in few iterations, which is far from enough for neural networks to overfit these queries. The underfitting usually results in severe drifts in camera tracking and artifacts in reconstruction. To resolve this issue, we propose query quantized neural SLAM which uses quantized queries to reduce variations of input for much easier and faster overfitting a frame. To this end, we quantize a query into a discrete representation with a set of codes, and only allow neural networks to observe a finite number of variations. This allows neural networks to become increasingly familiar with these codes after overfitting more and more previous frames. Moreover, we also introduce novel initialization, losses, and argumentation to stabilize the optimization with significant uncertainty in the early optimization stage, constrain the optimization space, and estimate camera poses more accurately. We justify the effectiveness of each design and report visual and numerical comparisons on widely used benchmarks to show our superiority over the latest methods in both reconstruction and camera tracking. △ Less

Submitted 20 December, 2024; originally announced December 2024.

Comments: To be appeared at AAAI25

Sensing Surface Patches in Volume Rendering for Inferring Signed Distance Functions

Authors: Sijia Jiang, Tong Wu, Jing Hua, Zhizhong Han

Abstract: It is vital to recover 3D geometry from multi-view RGB images in many 3D computer vision tasks. The latest methods infer the geometry represented as a signed distance field by minimizing the rendering error on the field through volume rendering. However, it is still challenging to explicitly impose constraints on surfaces for inferring more geometry details due to the limited ability of sensing su… ▽ More It is vital to recover 3D geometry from multi-view RGB images in many 3D computer vision tasks. The latest methods infer the geometry represented as a signed distance field by minimizing the rendering error on the field through volume rendering. However, it is still challenging to explicitly impose constraints on surfaces for inferring more geometry details due to the limited ability of sensing surfaces in volume rendering. To resolve this problem, we introduce a method to infer signed distance functions (SDFs) with a better sense of surfaces through volume rendering. Using the gradients and signed distances, we establish a small surface patch centered at the estimated intersection along a ray by pulling points randomly sampled nearby. Hence, we are able to explicitly impose surface constraints on the sensed surface patch, such as multi-view photo consistency and supervision from depth or normal priors, through volume rendering. We evaluate our method by numerical and visual comparisons on scene benchmarks. Our superiority over the latest methods justifies our effectiveness. △ Less

Submitted 20 December, 2024; originally announced December 2024.

Comments: To be appeared at AAAI25

EVLM: Self-Reflective Multimodal Reasoning for Cross-Dimensional Visual Editing

Authors: Umar Khalid, Kashif Munir, Hasan Iqbal, Azib Farooq, Jing Hua, Nazanin Rahnavard, Chen Chen, Victor Zhu, Zhengping Ji

Abstract: Editing complex visual content from ambiguous or partially specified instructions remains a core challenge in vision-language modeling. Existing models can contextualize content but often fail to infer the underlying intent within a reference image or scene, leading to inconsistent or misaligned edits. We introduce the Editing Vision-Language Model (EVLM), a system that interprets ambiguous instru… ▽ More Editing complex visual content from ambiguous or partially specified instructions remains a core challenge in vision-language modeling. Existing models can contextualize content but often fail to infer the underlying intent within a reference image or scene, leading to inconsistent or misaligned edits. We introduce the Editing Vision-Language Model (EVLM), a system that interprets ambiguous instructions in conjunction with reference visuals to produce precise, context-aware editing prompts. EVLM's key innovation is a reflective reasoning framework that translates subjective user intent into structured, actionable outputs by aligning with human-rated rationales through Reflection-Aware KL-Divergence Target Optimization (RKTO). By combining Chain-of-Thought (CoT) reasoning with RKTO alignment, EVLM captures fine-grained editing preferences without relying on binary supervision. Trained on a dataset of 30,000 CoT examples with human-annotated rationale quality, EVLM achieves substantial gains in alignment with human intent. Experiments across image, video, 3D, and 4D editing tasks show that EVLM generates coherent and high-quality instructions, providing a scalable foundation for multimodal editing and reasoning. △ Less

Submitted 7 November, 2025; v1 submitted 13 December, 2024; originally announced December 2024.

Comments: Technical Report

Cycle-Constrained Adversarial Denoising Convolutional Network for PET Image Denoising: Multi-Dimensional Validation on Large Datasets with Reader Study and Real Low-Dose Data

Authors: Yucun Hou, Fenglin Zhan, Xin Cheng, Chenxi Li, Ziquan Yuan, Runze Liao, Haihao Wang, Jianlang Hua, Jing Wu, Jianyong Jiang

Abstract: Positron emission tomography (PET) is a critical tool for diagnosing tumors and neurological disorders but poses radiation risks to patients, particularly to sensitive populations. While reducing injected radiation dose mitigates this risk, it often compromises image quality. To reconstruct full-dose-quality images from low-dose scans, we propose a Cycle-constrained Adversarial Denoising Convoluti… ▽ More Positron emission tomography (PET) is a critical tool for diagnosing tumors and neurological disorders but poses radiation risks to patients, particularly to sensitive populations. While reducing injected radiation dose mitigates this risk, it often compromises image quality. To reconstruct full-dose-quality images from low-dose scans, we propose a Cycle-constrained Adversarial Denoising Convolutional Network (Cycle-DCN). This model integrates a noise predictor, two discriminators, and a consistency network, and is optimized using a combination of supervised loss, adversarial loss, cycle consistency loss, identity loss, and neighboring Structural Similarity Index (SSIM) loss. Experiments were conducted on a large dataset consisting of raw PET brain data from 1,224 patients, acquired using a Siemens Biograph Vision PET/CT scanner. Each patient underwent a 120-seconds brain scan. To simulate low-dose PET conditions, images were reconstructed from shortened scan durations of 30, 12, and 5 seconds, corresponding to 1/4, 1/10, and 1/24 of the full-dose acquisition, respectively, using a custom-developed GPU-based image reconstruction software. The results show that Cycle-DCN significantly improves average Peak Signal-to-Noise Ratio (PSNR), SSIM, and Normalized Root Mean Square Error (NRMSE) across three dose levels, with improvements of up to 56%, 35%, and 71%, respectively. Additionally, it achieves contrast-to-noise ratio (CNR) and Edge Preservation Index (EPI) values that closely align with full-dose images, effectively preserving image details, tumor shape, and contrast, while resolving issues with blurred edges. The results of reader studies indicated that the images restored by Cycle-DCN consistently received the highest ratings from nuclear medicine physicians, highlighting their strong clinical relevance. △ Less

Submitted 31 October, 2024; originally announced October 2024.

Comments: This work has been submitted to the IEEE for possible publication

Journal ref: Med Image Anal. 107(Pt B) (2026) 103826

NASM: Neural Anisotropic Surface Meshing

Authors: Hongbo Li, Haikuan Zhu, Sikai Zhong, Ningna Wang, Cheng Lin, Xiaohu Guo, Shiqing Xin, Wenping Wang, Jing Hua, Zichun Zhong

Abstract: This paper introduces a new learning-based method, NASM, for anisotropic surface meshing. Our key idea is to propose a graph neural network to embed an input mesh into a high-dimensional (high-d) Euclidean embedding space to preserve curvature-based anisotropic metric by using a dot product loss between high-d edge vectors. This can dramatically reduce the computational time and increase the scala… ▽ More This paper introduces a new learning-based method, NASM, for anisotropic surface meshing. Our key idea is to propose a graph neural network to embed an input mesh into a high-dimensional (high-d) Euclidean embedding space to preserve curvature-based anisotropic metric by using a dot product loss between high-d edge vectors. This can dramatically reduce the computational time and increase the scalability. Then, we propose a novel feature-sensitive remeshing on the generated high-d embedding to automatically capture sharp geometric features. We define a high-d normal metric, and then derive an automatic differentiation on a high-d centroidal Voronoi tessellation (CVT) optimization with the normal metric to simultaneously preserve geometric features and curvature anisotropy that exhibit in the original 3D shapes. To our knowledge, this is the first time that a deep learning framework and a large dataset are proposed to construct a high-d Euclidean embedding space for 3D anisotropic surface meshing. Experimental results are evaluated and compared with the state-of-the-art in anisotropic surface meshing on a large number of surface models from Thingi10K dataset as well as tested on extensive unseen 3D shapes from Multi-Garment Network dataset and FAUST human dataset. △ Less

Submitted 31 October, 2024; v1 submitted 30 October, 2024; originally announced October 2024.

Comments: SIGGRAPH Asia 2024 (Conference Track)

GIG: Graph Data Imputation With Graph Differential Dependencies

Authors: Jiang Hua, Michael Bewong, Selasi Kwashie, MD Geaur Rahman, Junwei Hu, Xi Guo, Zaiwen Fen

Abstract: Data imputation addresses the challenge of imputing missing values in database instances, ensuring consistency with the overall semantics of the dataset. Although several heuristics which rely on statistical methods, and ad-hoc rules have been proposed. These do not generalise well and often lack data context. Consequently, they also lack explainability. The existing techniques also mostly focus o… ▽ More Data imputation addresses the challenge of imputing missing values in database instances, ensuring consistency with the overall semantics of the dataset. Although several heuristics which rely on statistical methods, and ad-hoc rules have been proposed. These do not generalise well and often lack data context. Consequently, they also lack explainability. The existing techniques also mostly focus on the relational data context making them unsuitable for wider application contexts such as in graph data. In this paper, we propose a graph data imputation approach called GIG which relies on graph differential dependencies (GDDs). GIG, learns the GDDs from a given knowledge graph, and uses these rules to train a transformer model which then predicts the value of missing data within the graph. By leveraging GDDs, GIG incoporates semantic knowledge into the data imputation process making it more reliable and explainable. Experimental results on seven real-world datasets highlight GIG's effectiveness compared to existing state-of-the-art approaches. △ Less

Submitted 21 October, 2024; originally announced October 2024.

Comments: 12 pages, 4 figures, published to ADC

A Survey on Computational Solutions for Reconstructing Complete Objects by Reassembling Their Fractured Parts

Authors: Jiaxin Lu, Yongqing Liang, Huijun Han, Jiacheng Hua, Junfeng Jiang, Xin Li, Qixing Huang

Abstract: Reconstructing a complete object from its parts is a fundamental problem in many scientific domains. The purpose of this article is to provide a systematic survey on this topic. The reassembly problem requires understanding the attributes of individual pieces and establishing matches between different pieces. Many approaches also model priors of the underlying complete object. Existing approaches… ▽ More Reconstructing a complete object from its parts is a fundamental problem in many scientific domains. The purpose of this article is to provide a systematic survey on this topic. The reassembly problem requires understanding the attributes of individual pieces and establishing matches between different pieces. Many approaches also model priors of the underlying complete object. Existing approaches are tightly connected problems of shape segmentation, shape matching, and learning shape priors. We provide existing algorithms in this context and emphasize their similarities and differences to general-purpose approaches. We also survey the trends from early non-deep learning approaches to more recent deep learning approaches. In addition to algorithms, this survey will also describe existing datasets, open-source software packages, and applications. To the best of our knowledge, this is the first comprehensive survey on this topic in computer graphics. △ Less

Submitted 27 March, 2025; v1 submitted 18 October, 2024; originally announced October 2024.

Comments: 36 pages, 22 figures