Showing 1–50 of 77 results for author: Shu, T

SimWorld-Robotics: Synthesizing Photorealistic and Dynamic Urban Environments for Multimodal Robot Navigation and Collaboration

Authors: Yan Zhuang, Jiawei Ren, Xiaokang Ye, Jianzhi Shen, Ruixuan Zhang, Tianai Yue, Muhammad Faayez, Xuhong He, Ziqiao Ma, Lianhui Qin, Zhiting Hu, Tianmin Shu

Abstract: Recent advances in foundation models have shown promising results in developing generalist robotics that can perform diverse tasks in open-ended scenarios given multimodal inputs. However, current work has been mainly focused on indoor, household scenarios. In this work, we present SimWorld-Robotics~(SWR), a simulation platform for embodied AI in large-scale, photorealistic urban environments. Bui… ▽ More Recent advances in foundation models have shown promising results in developing generalist robotics that can perform diverse tasks in open-ended scenarios given multimodal inputs. However, current work has been mainly focused on indoor, household scenarios. In this work, we present SimWorld-Robotics~(SWR), a simulation platform for embodied AI in large-scale, photorealistic urban environments. Built on Unreal Engine 5, SWR procedurally generates unlimited photorealistic urban scenes populated with dynamic elements such as pedestrians and traffic systems, surpassing prior urban simulations in realism, complexity, and scalability. It also supports multi-robot control and communication. With these key features, we build two challenging robot benchmarks: (1) a multimodal instruction-following task, where a robot must follow vision-language navigation instructions to reach a destination in the presence of pedestrians and traffic; and (2) a multi-agent search task, where two robots must communicate to cooperatively locate and meet each other. Unlike existing benchmarks, these two new benchmarks comprehensively evaluate a wide range of critical robot capacities in realistic scenarios, including (1) multimodal instructions grounding, (2) 3D spatial reasoning in large environments, (3) safe, long-range navigation with people and traffic, (4) multi-robot collaboration, and (5) grounded communication. Our experimental results demonstrate that state-of-the-art models, including vision-language models (VLMs), struggle with our tasks, lacking robust perception, reasoning, and planning abilities necessary for urban environments. △ Less

Submitted 10 December, 2025; originally announced December 2025.

Comments: Conference: NeurIPS 2025 (main)

SimWorld: An Open-ended Realistic Simulator for Autonomous Agents in Physical and Social Worlds

Authors: Jiawei Ren, Yan Zhuang, Xiaokang Ye, Lingjun Mao, Xuhong He, Jianzhi Shen, Mrinaal Dogra, Yiming Liang, Ruixuan Zhang, Tianai Yue, Yiqing Yang, Eric Liu, Ryan Wu, Kevin Benavente, Rajiv Mandya Nagaraju, Muhammad Faayez, Xiyan Zhang, Dhruv Vivek Sharma, Xianrui Zhong, Ziqiao Ma, Tianmin Shu, Zhiting Hu, Lianhui Qin

Abstract: While LLM/VLM-powered AI agents have advanced rapidly in math, coding, and computer use, their applications in complex physical and social environments remain challenging. Building agents that can survive and thrive in the real world (for example, by autonomously earning income or running a business) requires massive-scale interaction, reasoning, training, and evaluation across diverse embodied sc… ▽ More While LLM/VLM-powered AI agents have advanced rapidly in math, coding, and computer use, their applications in complex physical and social environments remain challenging. Building agents that can survive and thrive in the real world (for example, by autonomously earning income or running a business) requires massive-scale interaction, reasoning, training, and evaluation across diverse embodied scenarios. However, existing world simulators for such development fall short: they often rely on limited hand-crafted environments, simulate simplified game-like physics and social rules, and lack native support for LLM/VLM agents. We introduce SimWorld, a new simulator built on Unreal Engine 5, designed for developing and evaluating LLM/VLM agents in rich, real-world-like settings. SimWorld offers three core capabilities: (1) realistic, open-ended world simulation, including accurate physical and social dynamics and language-driven procedural environment generation; (2) a rich interface for LLM/VLM agents, with multimodal world inputs and open-vocabulary actions at varying levels of abstraction; and (3) diverse and extensible physical and social reasoning scenarios that are easily customizable by users. We demonstrate SimWorld by deploying frontier LLM agents (e.g., GPT-4o, Gemini-2.5-Flash, Claude-3.5, and DeepSeek-Prover-V2) on long-horizon multi-agent delivery tasks involving strategic cooperation and competition. The results reveal distinct reasoning patterns and limitations across models. We open-source SimWorld and hope it becomes a foundational platform for advancing real-world agent intelligence across disciplines: https://simworld.org. △ Less

Submitted 30 November, 2025; originally announced December 2025.

World-in-World: World Models in a Closed-Loop World

Authors: Jiahan Zhang, Muqing Jiang, Nanru Dai, Taiming Lu, Arda Uzunoglu, Shunchi Zhang, Yana Wei, Jiahao Wang, Vishal M. Patel, Paul Pu Liang, Daniel Khashabi, Cheng Peng, Rama Chellappa, Tianmin Shu, Alan Yuille, Yilun Du, Jieneng Chen

Abstract: Generative world models (WMs) can now simulate worlds with striking visual realism, which naturally raises the question of whether they can endow embodied agents with predictive perception for decision making. Progress on this question has been limited by fragmented evaluation: most existing benchmarks adopt open-loop protocols that emphasize visual quality in isolation, leaving the core issue of… ▽ More Generative world models (WMs) can now simulate worlds with striking visual realism, which naturally raises the question of whether they can endow embodied agents with predictive perception for decision making. Progress on this question has been limited by fragmented evaluation: most existing benchmarks adopt open-loop protocols that emphasize visual quality in isolation, leaving the core issue of embodied utility unresolved, i.e., do WMs actually help agents succeed at embodied tasks? To address this gap, we introduce World-in-World, the first open platform that benchmarks WMs in a closed-loop world that mirrors real agent-environment interactions. World-in-World provides a unified online planning strategy and a standardized action API, enabling heterogeneous WMs for decision making. We curate four closed-loop environments that rigorously evaluate diverse WMs, prioritize task success as the primary metric, and move beyond the common focus on visual quality; we also present the first data scaling law for world models in embodied settings. Our study uncovers three surprises: (1) visual quality alone does not guarantee task success, controllability matters more; (2) scaling post-training with action-observation data is more effective than upgrading the pretrained video generators; and (3) allocating more inference-time compute allows WMs to substantially improve closed-loop performance. △ Less

Submitted 20 October, 2025; originally announced October 2025.

Comments: Code is at https://github.com/World-In-World/world-in-world

DPSformer: A long-tail-aware model for improving heavy rainfall prediction

Authors: Zenghui Huang, Ting Shu, Zhonglei Wang, Yang Lu, Yan Yan, Wei Zhong, Hanzi Wang

Abstract: Accurate and timely forecasting of heavy rainfall remains a critical challenge for modern society. Precipitation exhibits a highly imbalanced distribution: most observations record no or light rain, while heavy rainfall events are rare. Such an imbalanced distribution obstructs deep learning models from effectively predicting heavy rainfall events. To address this challenge, we treat rainfall fore… ▽ More Accurate and timely forecasting of heavy rainfall remains a critical challenge for modern society. Precipitation exhibits a highly imbalanced distribution: most observations record no or light rain, while heavy rainfall events are rare. Such an imbalanced distribution obstructs deep learning models from effectively predicting heavy rainfall events. To address this challenge, we treat rainfall forecasting explicitly as a long-tailed learning problem, identifying the insufficient representation of heavy rainfall events as the primary barrier to forecasting accuracy. Therefore, we introduce DPSformer, a long-tail-aware model that enriches representation of heavy rainfall events through a high-resolution branch. For heavy rainfall events $ \geq $ 50 mm/6 h, DPSformer lifts the Critical Success Index (CSI) of a baseline Numerical Weather Prediction (NWP) model from 0.012 to 0.067. For the top 1% coverage of heavy rainfall events, its Fraction Skill Score (FSS) exceeds 0.45, surpassing existing methods. Our work establishes an effective long-tailed paradigm for heavy rainfall prediction, offering a practical tool to enhance early warning systems and mitigate the societal impacts of extreme weather events. △ Less

Submitted 20 September, 2025; originally announced September 2025.

The Era of Real-World Human Interaction: RL from User Conversations

Authors: Chuanyang Jin, Jing Xu, Bo Liu, Leitian Tao, Olga Golovneva, Tianmin Shu, Wenting Zhao, Xian Li, Jason Weston

Abstract: We posit that to achieve continual model improvement and multifaceted alignment, future models must learn from natural human interaction. Current conversational models are aligned using pre-annotated, expert-generated human feedback. In this work, we introduce Reinforcement Learning from Human Interaction (RLHI), a paradigm that learns directly from in-the-wild user conversations. We develop two c… ▽ More We posit that to achieve continual model improvement and multifaceted alignment, future models must learn from natural human interaction. Current conversational models are aligned using pre-annotated, expert-generated human feedback. In this work, we introduce Reinforcement Learning from Human Interaction (RLHI), a paradigm that learns directly from in-the-wild user conversations. We develop two complementary methods: (1) RLHI with User-Guided Rewrites, which revises unsatisfactory model outputs based on users' natural-language follow-up responses, (2) RLHI with User-Based Rewards, which learns via a reward model conditioned on knowledge of the user's long-term interaction history (termed persona). Together, these methods link long-term user personas to turn-level preferences via persona-conditioned preference optimization. Trained on conversations derived from WildChat, both RLHI variants outperform strong baselines in personalization and instruction-following, and similar feedback enhances performance on reasoning benchmarks. These results suggest organic human interaction offers scalable, effective supervision for personalized alignment. △ Less

Submitted 29 September, 2025; originally announced September 2025.

ProToM: Promoting Prosocial Behaviour via Theory of Mind-Informed Feedback

Authors: Matteo Bortoletto, Yichao Zhou, Lance Ying, Tianmin Shu, Andreas Bulling

Abstract: While humans are inherently social creatures, the challenge of identifying when and how to assist and collaborate with others - particularly when pursuing independent goals - can hinder cooperation. To address this challenge, we aim to develop an AI system that provides useful feedback to promote prosocial behaviour - actions that benefit others, even when not directly aligned with one's own goals… ▽ More While humans are inherently social creatures, the challenge of identifying when and how to assist and collaborate with others - particularly when pursuing independent goals - can hinder cooperation. To address this challenge, we aim to develop an AI system that provides useful feedback to promote prosocial behaviour - actions that benefit others, even when not directly aligned with one's own goals. We introduce ProToM, a Theory of Mind-informed facilitator that promotes prosocial actions in multi-agent systems by providing targeted, context-sensitive feedback to individual agents. ProToM first infers agents' goals using Bayesian inverse planning, then selects feedback to communicate by maximising expected utility, conditioned on the inferred goal distribution. We evaluate our approach against baselines in two multi-agent environments: Doors, Keys, and Gems, as well as Overcooked. Our results suggest that state-of-the-art large language and reasoning models fall short of communicating feedback that is both contextually grounded and well-timed - leading to higher communication overhead and task speedup. In contrast, ProToM provides targeted and helpful feedback, achieving a higher success rate, shorter task completion times, and is consistently preferred by human users. △ Less

Submitted 5 September, 2025; originally announced September 2025.

Comments: Website at https://www.matteobortoletto.org/protom/

Augmented Vision-Language Models: A Systematic Review

Authors: Anthony C Davis, Burhan Sadiq, Tianmin Shu, Chien-Ming Huang

Abstract: Recent advances in visual-language machine learning models have demonstrated exceptional ability to use natural language and understand visual scenes by training on large, unstructured datasets. However, this training paradigm cannot produce interpretable explanations for its outputs, requires retraining to integrate new information, is highly resource-intensive, and struggles with certain forms o… ▽ More Recent advances in visual-language machine learning models have demonstrated exceptional ability to use natural language and understand visual scenes by training on large, unstructured datasets. However, this training paradigm cannot produce interpretable explanations for its outputs, requires retraining to integrate new information, is highly resource-intensive, and struggles with certain forms of logical reasoning. One promising solution involves integrating neural networks with external symbolic information systems, forming neural symbolic systems that can enhance reasoning and memory abilities. These neural symbolic systems provide more interpretable explanations to their outputs and the capacity to assimilate new information without extensive retraining. Utilizing powerful pre-trained Vision-Language Models (VLMs) as the core neural component, augmented by external systems, offers a pragmatic approach to realizing the benefits of neural-symbolic integration. This systematic literature review aims to categorize techniques through which visual-language understanding can be improved by interacting with external symbolic information systems. △ Less

Submitted 24 July, 2025; originally announced July 2025.

Do Vision-Language Models Have Internal World Models? Towards an Atomic Evaluation

Authors: Qiyue Gao, Xinyu Pi, Kevin Liu, Junrong Chen, Ruolan Yang, Xinqi Huang, Xinyu Fang, Lu Sun, Gautham Kishore, Bo Ai, Stone Tao, Mengyang Liu, Jiaxi Yang, Chao-Jung Lai, Chuanyang Jin, Jiannan Xiang, Benhao Huang, Zeming Chen, David Danks, Hao Su, Tianmin Shu, Ziqiao Ma, Lianhui Qin, Zhiting Hu

Abstract: Internal world models (WMs) enable agents to understand the world's state and predict transitions, serving as the basis for advanced deliberative reasoning. Recent large Vision-Language Models (VLMs), such as OpenAI o3, GPT-4o and Gemini, exhibit potential as general-purpose WMs. While the latest studies have evaluated and shown limitations in specific capabilities such as visual understanding, a… ▽ More Internal world models (WMs) enable agents to understand the world's state and predict transitions, serving as the basis for advanced deliberative reasoning. Recent large Vision-Language Models (VLMs), such as OpenAI o3, GPT-4o and Gemini, exhibit potential as general-purpose WMs. While the latest studies have evaluated and shown limitations in specific capabilities such as visual understanding, a systematic evaluation of VLMs' fundamental WM abilities remains absent. Drawing on comparative psychology and cognitive science, we propose a two-stage framework that assesses Perception (visual, spatial, temporal, quantitative, and motion) and Prediction (mechanistic simulation, transitive inference, compositional inference) to provide an atomic evaluation of VLMs as WMs. Guided by this framework, we introduce WM-ABench, a large-scale benchmark comprising 23 fine-grained evaluation dimensions across 6 diverse simulated environments with controlled counterfactual simulations. Through 660 experiments on 15 latest commercial and open-source VLMs, we find that these models exhibit striking limitations in basic world modeling abilities. For instance, almost all models perform at near-random accuracy when distinguishing motion trajectories. Additionally, they lack disentangled understanding -- e.g., some models tend to believe blue objects move faster than green ones. More rich results and analyses reveal significant gaps between VLMs and human-level world modeling. △ Less

Submitted 26 June, 2025; originally announced June 2025.

Comments: ACL 2025 (Findings)

PartInstruct: Part-level Instruction Following for Fine-grained Robot Manipulation

Authors: Yifan Yin, Zhengtao Han, Shivam Aarya, Jianxin Wang, Shuhang Xu, Jiawei Peng, Angtian Wang, Alan Yuille, Tianmin Shu

Abstract: Fine-grained robot manipulation, such as lifting and rotating a bottle to display the label on the cap, requires robust reasoning about object parts and their relationships with intended tasks. Despite recent advances in training general-purpose robot manipulation policies guided by language instructions, there is a notable lack of large-scale datasets for fine-grained manipulation tasks with part… ▽ More Fine-grained robot manipulation, such as lifting and rotating a bottle to display the label on the cap, requires robust reasoning about object parts and their relationships with intended tasks. Despite recent advances in training general-purpose robot manipulation policies guided by language instructions, there is a notable lack of large-scale datasets for fine-grained manipulation tasks with part-level instructions and diverse 3D object instances annotated with part-level labels. In this work, we introduce PartInstruct, the first large-scale benchmark for training and evaluating fine-grained robot manipulation models using part-level instructions. PartInstruct comprises 513 object instances across 14 categories, each annotated with part-level information, and 1302 fine-grained manipulation tasks organized into 16 task classes. Our training set consists of over 10,000 expert demonstrations synthesized in a 3D simulator, where each demonstration is paired with a high-level task instruction, a chain of base part-based skill instructions, and ground-truth 3D information about the object and its parts. Additionally, we designed a comprehensive test suite to evaluate the generalizability of learned policies across new states, objects, and tasks. We evaluated several state-of-the-art robot manipulation approaches, including end-to-end vision-language policy learning and bi-level planning models for robot manipulation on our benchmark. The experimental results reveal that current models struggle to robustly ground part concepts and predict actions in 3D space, and face challenges when manipulating object parts in long-horizon tasks. △ Less

Submitted 16 June, 2025; v1 submitted 27 May, 2025; originally announced May 2025.

Direct Preference Optimization for Adaptive Concept-based Explanations

Authors: Jacopo Teneggi, Zhenzhen Wang, Paul H. Yi, Tianmin Shu, Jeremias Sulam

Abstract: Concept-based explanation methods aim at making machine learning models more transparent by finding the most important semantic features of an input (e.g., colors, patterns, shapes) for a given prediction task. However, these methods generally ignore the communicative context of explanations, such as the preferences of a listener. For example, medical doctors understand explanations in terms of cl… ▽ More Concept-based explanation methods aim at making machine learning models more transparent by finding the most important semantic features of an input (e.g., colors, patterns, shapes) for a given prediction task. However, these methods generally ignore the communicative context of explanations, such as the preferences of a listener. For example, medical doctors understand explanations in terms of clinical markers, but patients may not, needing a different vocabulary to rationalize the same diagnosis. We address this gap with listener-adaptive explanations grounded in principles of pragmatic reasoning and the rational speech act. We introduce an iterative training procedure based on direct preference optimization where a speaker learns to compose explanations that maximize communicative utility for a listener. Our approach only needs access to pairwise preferences, which can be collected from human feedback, making it particularly relevant in real-world scenarios where a model of the listener may not be available. We demonstrate that our method is able to align speakers with the preferences of simulated listeners on image classification across three datasets, and further validate that pragmatic explanations generated with our method improve the classification accuracy of participants in a user study. △ Less

Submitted 1 October, 2025; v1 submitted 21 May, 2025; originally announced May 2025.

Position: Foundation Models Need Digital Twin Representations

Authors: Yiqing Shen, Hao Ding, Lalithkumar Seenivasan, Tianmin Shu, Mathias Unberath

Abstract: Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing… ▽ More Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing fine-grained spatial-temporal dynamics, and performing causal reasoning. These limitations cannot be overcome by simply scaling up model size or expanding datasets. This position paper argues that the machine learning community should consider digital twin (DT) representations, which are outcome-driven digital representations that serve as building blocks for creating virtual replicas of physical processes, as an alternative to the token representation for building FMs. Finally, we discuss how DT representations can address these challenges by providing physically grounded representations that explicitly encode domain knowledge and preserve the continuous nature of real-world processes. △ Less

Submitted 1 May, 2025; originally announced May 2025.

RealWebAssist: A Benchmark for Long-Horizon Web Assistance with Real-World Users

Authors: Suyu Ye, Haojun Shi, Darren Shih, Hyokun Yun, Tanya Roosta, Tianmin Shu

Abstract: To achieve successful assistance with long-horizon web-based tasks, AI agents must be able to sequentially follow real-world user instructions over a long period. Unlike existing web-based agent benchmarks, sequential instruction following in the real world poses significant challenges beyond performing a single, clearly defined task. For instance, real-world human instructions can be ambiguous, r… ▽ More To achieve successful assistance with long-horizon web-based tasks, AI agents must be able to sequentially follow real-world user instructions over a long period. Unlike existing web-based agent benchmarks, sequential instruction following in the real world poses significant challenges beyond performing a single, clearly defined task. For instance, real-world human instructions can be ambiguous, require different levels of AI assistance, and may evolve over time, reflecting changes in the user's mental state. To address this gap, we introduce RealWebAssist, a novel benchmark designed to evaluate sequential instruction-following in realistic scenarios involving long-horizon interactions with the web, visual GUI grounding, and understanding ambiguous real-world user instructions. RealWebAssist includes a dataset of sequential instructions collected from real-world human users. Each user instructs a web-based assistant to perform a series of tasks on multiple websites. A successful agent must reason about the true intent behind each instruction, keep track of the mental state of the user, understand user-specific routines, and ground the intended tasks to actions on the correct GUI elements. Our experimental results show that state-of-the-art models struggle to understand and ground user instructions, posing critical challenges in following real-world user instructions for long-horizon web assistance. △ Less

Submitted 1 December, 2025; v1 submitted 14 April, 2025; originally announced April 2025.

Comments: Project Website: https://scai.cs.jhu.edu/projects/RealWebAssist/ Code: https://github.com/SCAI-JHU/RealWebAssist

FADConv: A Frequency-Aware Dynamic Convolution for Farmland Non-agriculturalization Identification and Segmentation

Authors: Tan Shu, Li Shen

Abstract: Cropland non-agriculturalization refers to the conversion of arable land into non-agricultural uses such as forests, residential areas, and construction sites. This phenomenon not only directly leads to the loss of cropland resources but also poses systemic threats to food security and agricultural sustainability. Accurate identification of cropland and non-cropland areas is crucial for detecting… ▽ More Cropland non-agriculturalization refers to the conversion of arable land into non-agricultural uses such as forests, residential areas, and construction sites. This phenomenon not only directly leads to the loss of cropland resources but also poses systemic threats to food security and agricultural sustainability. Accurate identification of cropland and non-cropland areas is crucial for detecting and addressing this issue. Traditional CNNs employ static convolution layers, while dynamic convolution studies demonstrate that adaptively weighting multiple convolutional kernels through attention mechanisms can enhance accuracy. However, existing dynamic convolution methods relying on Global Average Pooling (GAP) for attention weight allocation suffer from information loss, limiting segmentation precision. This paper proposes Frequency-Aware Dynamic Convolution (FADConv) and a Frequency Attention (FAT) module to address these limitations. Building upon the foundational structure of dynamic convolution, we designed FADConv by integrating 2D Discrete Cosine Transform (2D DCT) to capture frequency domain features and fuse them. FAT module generates high-quality attention weights that replace the traditional GAP method,making the combination between dynamic convolution kernels more reasonable.Experiments on the GID and Hi-CNA datasets demonstrate that FADConv significantly improves segmentation accuracy with minimal computational overhead. For instance, ResNet18 with FADConv achieves 1.9% and 2.7% increases in F1-score and IoU for cropland segmentation on GID, with only 58.87M additional MAdds. Compared to other dynamic convolution approaches, FADConv exhibits superior performance in cropland segmentation tasks. △ Less

Submitted 4 April, 2025; originally announced April 2025.

On Benchmarking Human-Like Intelligence in Machines

Authors: Lance Ying, Katherine M. Collins, Lionel Wong, Ilia Sucholutsky, Ryan Liu, Adrian Weller, Tianmin Shu, Thomas L. Griffiths, Joshua B. Tenenbaum

Abstract: Recent benchmark studies have claimed that AI has approached or even surpassed human-level performances on various cognitive tasks. However, this position paper argues that current AI evaluation paradigms are insufficient for assessing human-like cognitive capabilities. We identify a set of key shortcomings: a lack of human-validated labels, inadequate representation of human response variability… ▽ More Recent benchmark studies have claimed that AI has approached or even surpassed human-level performances on various cognitive tasks. However, this position paper argues that current AI evaluation paradigms are insufficient for assessing human-like cognitive capabilities. We identify a set of key shortcomings: a lack of human-validated labels, inadequate representation of human response variability and uncertainty, and reliance on simplified and ecologically-invalid tasks. We support our claims by conducting a human evaluation study on ten existing AI benchmarks, suggesting significant biases and flaws in task and label designs. To address these limitations, we propose five concrete recommendations for developing future benchmarks that will enable more rigorous and meaningful evaluations of human-like cognitive capacities in AI with various implications for such AI applications. △ Less

Submitted 27 February, 2025; originally announced February 2025.

Comments: 18 pages, 5 figures

AutoToM: Scaling Model-based Mental Inference via Automated Agent Modeling

Authors: Zhining Zhang, Chuanyang Jin, Mung Yao Jia, Shunchi Zhang, Tianmin Shu

Abstract: Theory of Mind (ToM), the ability to understand people's minds based on their behavior, is key to developing socially intelligent agents. Current approaches to ToM reasoning either rely on prompting Large Language Models (LLMs), which are prone to systematic errors, or use handcrafted, rigid agent models for model-based inference, which are more robust but fail to generalize across domains. In thi… ▽ More Theory of Mind (ToM), the ability to understand people's minds based on their behavior, is key to developing socially intelligent agents. Current approaches to ToM reasoning either rely on prompting Large Language Models (LLMs), which are prone to systematic errors, or use handcrafted, rigid agent models for model-based inference, which are more robust but fail to generalize across domains. In this work, we introduce AutoToM, an automated agent modeling method for scalable, robust, and interpretable mental inference. Given a ToM problem, AutoToM first proposes an initial agent model and then performs automated Bayesian inverse planning based on this model, leveraging an LLM backend. Guided by inference uncertainty, it iteratively refines the model by introducing additional mental variables and/or incorporating more timesteps in the context. Across five diverse benchmarks, AutoToM outperforms existing ToM methods and even large reasoning models. Additionally, we show that AutoToM can produce human-like confidence estimates and enable online mental inference for embodied decision-making. △ Less

Submitted 29 June, 2025; v1 submitted 21 February, 2025; originally announced February 2025.

Comments: 39 pages, 10 figures, 13 tables. Website at https://chuanyangjin.com/AutoToM/

GenEx: Generating an Explorable World

Authors: Taiming Lu, Tianmin Shu, Junfei Xiao, Luoxin Ye, Jiahao Wang, Cheng Peng, Chen Wei, Daniel Khashabi, Rama Chellappa, Alan Yuille, Jieneng Chen

Abstract: Understanding, navigating, and exploring the 3D physical real world has long been a central challenge in the development of artificial intelligence. In this work, we take a step toward this goal by introducing GenEx, a system capable of planning complex embodied world exploration, guided by its generative imagination that forms priors (expectations) about the surrounding environments. GenEx genera… ▽ More Understanding, navigating, and exploring the 3D physical real world has long been a central challenge in the development of artificial intelligence. In this work, we take a step toward this goal by introducing GenEx, a system capable of planning complex embodied world exploration, guided by its generative imagination that forms priors (expectations) about the surrounding environments. GenEx generates an entire 3D-consistent imaginative environment from as little as a single RGB image, bringing it to life through panoramic video streams. Leveraging scalable 3D world data curated from Unreal Engine, our generative model is rounded in the physical world. It captures a continuous 360-degree environment with little effort, offering a boundless landscape for AI agents to explore and interact with. GenEx achieves high-quality world generation, robust loop consistency over long trajectories, and demonstrates strong 3D capabilities such as consistency and active 3D mapping. Powered by generative imagination of the world, GPT-assisted agents are equipped to perform complex embodied tasks, including both goal-agnostic exploration and goal-driven navigation. These agents utilize predictive expectation regarding unseen parts of the physical world to refine their beliefs, simulate different outcomes based on potential decisions, and make more informed choices. In summary, we demonstrate that GenEx provides a transformative platform for advancing embodied AI in imaginative spaces and brings potential for extending these capabilities to real-world exploration. △ Less

Submitted 20 January, 2025; v1 submitted 12 December, 2024; originally announced December 2024.

Comments: Website: GenEx.world

Generative World Explorer

Authors: Taiming Lu, Tianmin Shu, Alan Yuille, Daniel Khashabi, Jieneng Chen

Abstract: Planning with partial observation is a central challenge in embodied AI. A majority of prior works have tackled this challenge by developing agents that physically explore their environment to update their beliefs about the world state. In contrast, humans can $\textit{imagine}$ unseen parts of the world through a mental exploration and $\textit{revise}$ their beliefs with imagined observations. S… ▽ More Planning with partial observation is a central challenge in embodied AI. A majority of prior works have tackled this challenge by developing agents that physically explore their environment to update their beliefs about the world state. In contrast, humans can $\textit{imagine}$ unseen parts of the world through a mental exploration and $\textit{revise}$ their beliefs with imagined observations. Such updated beliefs can allow them to make more informed decisions, without necessitating the physical exploration of the world at all times. To achieve this human-like ability, we introduce the $\textit{Generative World Explorer (Genex)}$, an egocentric world exploration framework that allows an agent to mentally explore a large-scale 3D world (e.g., urban scenes) and acquire imagined observations to update its belief. This updated belief will then help the agent to make a more informed decision at the current step. To train $\textit{Genex}$, we create a synthetic urban scene dataset, Genex-DB. Our experimental results demonstrate that (1) $\textit{Genex}$ can generate high-quality and consistent observations during long-horizon exploration of a large virtual physical world and (2) the beliefs updated with the generated observations can inform an existing decision-making model (e.g., an LLM agent) to make better plans. △ Less

Submitted 8 September, 2025; v1 submitted 18 November, 2024; originally announced November 2024.

Comments: Website: generative-world-explorer.github.io

Few-Shot Task Learning through Inverse Generative Modeling

Authors: Aviv Netanyahu, Yilun Du, Antonia Bronars, Jyothish Pari, Joshua Tenenbaum, Tianmin Shu, Pulkit Agrawal

Abstract: Learning the intents of an agent, defined by its goals or motion style, is often extremely challenging from just a few examples. We refer to this problem as task concept learning and present our approach, Few-Shot Task Learning through Inverse Generative Modeling (FTL-IGM), which learns new task concepts by leveraging invertible neural generative models. The core idea is to pretrain a generative m… ▽ More Learning the intents of an agent, defined by its goals or motion style, is often extremely challenging from just a few examples. We refer to this problem as task concept learning and present our approach, Few-Shot Task Learning through Inverse Generative Modeling (FTL-IGM), which learns new task concepts by leveraging invertible neural generative models. The core idea is to pretrain a generative model on a set of basic concepts and their demonstrations. Then, given a few demonstrations of a new concept (such as a new goal or a new action), our method learns the underlying concepts through backpropagation without updating the model weights, thanks to the invertibility of the generative model. We evaluate our method in five domains -- object rearrangement, goal-oriented navigation, motion caption of human actions, autonomous driving, and real-world table-top manipulation. Our experimental results demonstrate that via the pretrained generative model, we successfully learn novel concepts and generate agent plans or motion corresponding to these concepts in (1) unseen environments and (2) in composition with training concepts. △ Less

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

Comments: Added acknowledgment

Constrained Human-AI Cooperation: An Inclusive Embodied Social Intelligence Challenge

Authors: Weihua Du, Qiushi Lyu, Jiaming Shan, Zhenting Qi, Hongxin Zhang, Sunli Chen, Andi Peng, Tianmin Shu, Kwonjoon Lee, Behzad Dariush, Chuang Gan

Abstract: We introduce Constrained Human-AI Cooperation (CHAIC), an inclusive embodied social intelligence challenge designed to test social perception and cooperation in embodied agents. In CHAIC, the goal is for an embodied agent equipped with egocentric observations to assist a human who may be operating under physical constraints -- e.g., unable to reach high places or confined to a wheelchair -- in per… ▽ More We introduce Constrained Human-AI Cooperation (CHAIC), an inclusive embodied social intelligence challenge designed to test social perception and cooperation in embodied agents. In CHAIC, the goal is for an embodied agent equipped with egocentric observations to assist a human who may be operating under physical constraints -- e.g., unable to reach high places or confined to a wheelchair -- in performing common household or outdoor tasks as efficiently as possible. To achieve this, a successful helper must: (1) infer the human's intents and constraints by following the human and observing their behaviors (social perception), and (2) make a cooperative plan tailored to the human partner to solve the task as quickly as possible, working together as a team (cooperative planning). To benchmark this challenge, we create four new agents with real physical constraints and eight long-horizon tasks featuring both indoor and outdoor scenes with various constraints, emergency events, and potential risks. We benchmark planning- and learning-based baselines on the challenge and introduce a new method that leverages large language models and behavior modeling. Empirical evaluations demonstrate the effectiveness of our benchmark in enabling systematic assessment of key aspects of machine social intelligence. Our benchmark and code are publicly available at https://github.com/UMass-Embodied-AGI/CHAIC. △ Less

Submitted 11 June, 2025; v1 submitted 3 November, 2024; originally announced November 2024.

Comments: NeurIPS 2024 Dataset and Benchmark Track. The first two authors contributed equally. Project Website at https://umass-embodied-agi.github.io/CHAIC/

Slide-based Graph Collaborative Training for Histopathology Whole Slide Image Analysis

Authors: Jun Shi, Tong Shu, Zhiguo Jiang, Wei Wang, Haibo Wu, Yushan Zheng

Abstract: The development of computational pathology lies in the consensus that pathological characteristics of tumors are significant guidance for cancer diagnostics. Most existing research focuses on the inner-contextual information within each WSI yet ignores the possible inter-correlations between slides. As the development of tumors is a continuous process involving a series of histological, morphologi… ▽ More The development of computational pathology lies in the consensus that pathological characteristics of tumors are significant guidance for cancer diagnostics. Most existing research focuses on the inner-contextual information within each WSI yet ignores the possible inter-correlations between slides. As the development of tumors is a continuous process involving a series of histological, morphological, and genetic changes that accumulate over time, the similarities and differences between WSIs across various stages, grades, locations and patients should potentially contribute to the representation of WSIs and deserve to be taken into account in WSI modeling. To verify the advancement of introducing the slide inter-correlations into the representation learning of WSIs, we proposed a generic WSI analysis pipeline SlideGCD that can be adapted to any existing Multiple Instance Learning (MIL) frameworks and improve their performance. With the new paradigm, the prior knowledge of cancer development can participate in the end-to-end workflow, which concurrently initializes and refines the slide representation, as a guide for message passing in the slide-based graph. Extensive comparisons and experiments are conducted to validate the effectiveness and robustness of the proposed pipeline across 4 different tasks, including cancer subtyping, cancer staging, survival prediction, and gene mutation prediction, with 7 representative SOTA WSI analysis frameworks as backbones. △ Less

Submitted 14 October, 2024; originally announced October 2024.

Pragmatic Embodied Spoken Instruction Following in Human-Robot Collaboration with Theory of Mind

Authors: Lance Ying, Xinyi Li, Shivam Aarya, Yizirui Fang, Yifan Yin, Jason Xinyu Liu, Stefanie Tellex, Joshua B. Tenenbaum, Tianmin Shu

Abstract: Spoken language instructions are ubiquitous in agent collaboration. However, in real-world human-robot collaboration, following human spoken instructions can be challenging due to various speaker and environmental factors, such as background noise or mispronunciation. When faced with noisy auditory inputs, humans can leverage the collaborative context in the embodied environment to interpret noisy… ▽ More Spoken language instructions are ubiquitous in agent collaboration. However, in real-world human-robot collaboration, following human spoken instructions can be challenging due to various speaker and environmental factors, such as background noise or mispronunciation. When faced with noisy auditory inputs, humans can leverage the collaborative context in the embodied environment to interpret noisy spoken instructions and take pragmatic assistive actions. In this paper, we present a cognitively inspired neurosymbolic model, Spoken Instruction Following through Theory of Mind (SIFToM), which leverages a Vision-Language Model with model-based mental inference to enable robots to pragmatically follow human instructions under diverse speech conditions. We test SIFToM in both simulated environments (VirtualHome) and real-world human-robot collaborative settings with human evaluations. Results show that SIFToM can significantly improve the performance of a lightweight base VLM (Gemini 2.5 Flash), outperforming state-of-the-art VLMs (Gemini 2.5 Pro) and approaching human-level accuracy on challenging spoken instruction following tasks. △ Less

Submitted 6 October, 2025; v1 submitted 16 September, 2024; originally announced September 2024.

Comments: 8 pages, 7 figures

MuMA-ToM: Multi-modal Multi-Agent Theory of Mind

Authors: Haojun Shi, Suyu Ye, Xinyu Fang, Chuanyang Jin, Leyla Isik, Yen-Ling Kuo, Tianmin Shu

Abstract: Understanding people's social interactions in complex real-world scenarios often relies on intricate mental reasoning. To truly understand how and why people interact with one another, we must infer the underlying mental states that give rise to the social interactions, i.e., Theory of Mind reasoning in multi-agent interactions. Additionally, social interactions are often multi-modal -- we can wat… ▽ More Understanding people's social interactions in complex real-world scenarios often relies on intricate mental reasoning. To truly understand how and why people interact with one another, we must infer the underlying mental states that give rise to the social interactions, i.e., Theory of Mind reasoning in multi-agent interactions. Additionally, social interactions are often multi-modal -- we can watch people's actions, hear their conversations, and/or read about their past behaviors. For AI systems to successfully and safely interact with people in real-world environments, they also need to understand people's mental states as well as their inferences about each other's mental states based on multi-modal information about their interactions. For this, we introduce MuMA-ToM, a Multi-modal Multi-Agent Theory of Mind benchmark. MuMA-ToM is the first multi-modal Theory of Mind benchmark that evaluates mental reasoning in embodied multi-agent interactions. In MuMA-ToM, we provide video and text descriptions of people's multi-modal behavior in realistic household environments. Based on the context, we then ask questions about people's goals, beliefs, and beliefs about others' goals. We validated MuMA-ToM in a human experiment and provided a human baseline. We also proposed a novel multi-modal, multi-agent ToM model, LIMP (Language model-based Inverse Multi-agent Planning). Our experimental results show that LIMP significantly outperforms state-of-the-art methods, including large multi-modal models (e.g., GPT-4o, Gemini-1.5 Pro) and a recent multi-modal ToM model, BIP-ALM. △ Less

Submitted 23 January, 2025; v1 submitted 22 August, 2024; originally announced August 2024.

Comments: AAAI-25 (Oral). Project website: https://scai.cs.jhu.edu/projects/MuMA-ToM/ Code: https://github.com/SCAI-JHU/MuMA-ToM

SlideGCD: Slide-based Graph Collaborative Training with Knowledge Distillation for Whole Slide Image Classification

Authors: Tong Shu, Jun Shi, Dongdong Sun, Zhiguo Jiang, Yushan Zheng

Abstract: Existing WSI analysis methods lie on the consensus that histopathological characteristics of tumors are significant guidance for cancer diagnostics. Particularly, as the evolution of cancers is a continuous process, the correlations and differences across various stages, anatomical locations and patients should be taken into account. However, recent research mainly focuses on the inner-contextual… ▽ More Existing WSI analysis methods lie on the consensus that histopathological characteristics of tumors are significant guidance for cancer diagnostics. Particularly, as the evolution of cancers is a continuous process, the correlations and differences across various stages, anatomical locations and patients should be taken into account. However, recent research mainly focuses on the inner-contextual information in a single WSI, ignoring the correlations between slides. To verify whether introducing the slide inter-correlations can bring improvements to WSI representation learning, we propose a generic WSI analysis pipeline SlideGCD that considers the existing multi-instance learning (MIL) methods as the backbone and forge the WSI classification task as a node classification problem. More specifically, SlideGCD declares a node buffer that stores previous slide embeddings for subsequent extensive slide-based graph construction and conducts graph learning to explore the inter-correlations implied in the slide-based graph. Moreover, we frame the MIL classifier and graph learning into two parallel workflows and deploy the knowledge distillation to transfer the differentiable information to the graph neural network. The consistent performance boosting, brought by SlideGCD, of four previous state-of-the-art MIL methods is observed on two TCGA benchmark datasets. The code is available at https://github.com/HFUT-miaLab/SlideGCD. △ Less

Submitted 19 July, 2024; v1 submitted 11 July, 2024; originally announced July 2024.

Comments: Accepted for MICCAI 2024

Learning Pareto Set for Multi-Objective Continuous Robot Control

Authors: Tianye Shu, Ke Shang, Cheng Gong, Yang Nan, Hisao Ishibuchi

Abstract: For a control problem with multiple conflicting objectives, there exists a set of Pareto-optimal policies called the Pareto set instead of a single optimal policy. When a multi-objective control problem is continuous and complex, traditional multi-objective reinforcement learning (MORL) algorithms search for many Pareto-optimal deep policies to approximate the Pareto set, which is quite resource-c… ▽ More For a control problem with multiple conflicting objectives, there exists a set of Pareto-optimal policies called the Pareto set instead of a single optimal policy. When a multi-objective control problem is continuous and complex, traditional multi-objective reinforcement learning (MORL) algorithms search for many Pareto-optimal deep policies to approximate the Pareto set, which is quite resource-consuming. In this paper, we propose a simple and resource-efficient MORL algorithm that learns a continuous representation of the Pareto set in a high-dimensional policy parameter space using a single hypernet. The learned hypernet can directly generate various well-trained policy networks for different user preferences. We compare our method with two state-of-the-art MORL algorithms on seven multi-objective continuous robot control problems. Experimental results show that our method achieves the best overall performance with the least training parameters. An interesting observation is that the Pareto set is well approximated by a curved line or surface in a high-dimensional parameter space. This observation will provide insight for researchers to design new MORL algorithms. △ Less

Submitted 27 June, 2024; originally announced June 2024.

Pragmatic Feature Preferences: Learning Reward-Relevant Preferences from Human Input

Authors: Andi Peng, Yuying Sun, Tianmin Shu, David Abel

Abstract: Humans use social context to specify preferences over behaviors, i.e. their reward functions. Yet, algorithms for inferring reward models from preference data do not take this social learning view into account. Inspired by pragmatic human communication, we study how to extract fine-grained data regarding why an example is preferred that is useful for learning more accurate reward models. We propos… ▽ More Humans use social context to specify preferences over behaviors, i.e. their reward functions. Yet, algorithms for inferring reward models from preference data do not take this social learning view into account. Inspired by pragmatic human communication, we study how to extract fine-grained data regarding why an example is preferred that is useful for learning more accurate reward models. We propose to enrich binary preference queries to ask both (1) which features of a given example are preferable in addition to (2) comparisons between examples themselves. We derive an approach for learning from these feature-level preferences, both for cases where users specify which features are reward-relevant, and when users do not. We evaluate our approach on linear bandit settings in both vision- and language-based domains. Results support the efficiency of our approach in quickly converging to accurate rewards with fewer comparisons vs. example-only labels. Finally, we validate the real-world applicability with a behavioral experiment on a mushroom foraging task. Our findings suggest that incorporating pragmatic feature preferences is a promising approach for more efficient user-aligned reward learning. △ Less

Submitted 23 May, 2024; originally announced May 2024.

Comments: ICML 2024

COMBO: Compositional World Models for Embodied Multi-Agent Cooperation

Authors: Hongxin Zhang, Zeyuan Wang, Qiushi Lyu, Zheyuan Zhang, Sunli Chen, Tianmin Shu, Behzad Dariush, Kwonjoon Lee, Yilun Du, Chuang Gan

Abstract: In this paper, we investigate the problem of embodied multi-agent cooperation, where decentralized agents must cooperate given only egocentric views of the world. To effectively plan in this setting, in contrast to learning world dynamics in a single-agent scenario, we must simulate world dynamics conditioned on an arbitrary number of agents' actions given only partial egocentric visual observatio… ▽ More In this paper, we investigate the problem of embodied multi-agent cooperation, where decentralized agents must cooperate given only egocentric views of the world. To effectively plan in this setting, in contrast to learning world dynamics in a single-agent scenario, we must simulate world dynamics conditioned on an arbitrary number of agents' actions given only partial egocentric visual observations of the world. To address this issue of partial observability, we first train generative models to estimate the overall world state given partial egocentric observations. To enable accurate simulation of multiple sets of actions on this world state, we then propose to learn a compositional world model for multi-agent cooperation by factorizing the naturally composable joint actions of multiple agents and compositionally generating the video conditioned on the world state. By leveraging this compositional world model, in combination with Vision Language Models to infer the actions of other agents, we can use a tree search procedure to integrate these modules and facilitate online cooperative planning. We evaluate our methods on three challenging benchmarks with 2-4 agents. The results show our compositional world model is effective and the framework enables the embodied agents to cooperate efficiently with different agents across various tasks and an arbitrary number of agents, showing the promising future of our proposed methods. More videos can be found at https://umass-embodied-agi.github.io/COMBO/. △ Less

Submitted 15 April, 2025; v1 submitted 16 April, 2024; originally announced April 2024.

Comments: Published at ICLR 2025. 24 pages. The first three authors contributed equally

Survey on Large Language Model-Enhanced Reinforcement Learning: Concept, Taxonomy, and Methods

Authors: Yuji Cao, Huan Zhao, Yuheng Cheng, Ting Shu, Yue Chen, Guolong Liu, Gaoqi Liang, Junhua Zhao, Jinyue Yan, Yun Li

Abstract: With extensive pre-trained knowledge and high-level general capabilities, large language models (LLMs) emerge as a promising avenue to augment reinforcement learning (RL) in aspects such as multi-task learning, sample efficiency, and high-level task planning. In this survey, we provide a comprehensive review of the existing literature in LLM-enhanced RL and summarize its characteristics compared t… ▽ More With extensive pre-trained knowledge and high-level general capabilities, large language models (LLMs) emerge as a promising avenue to augment reinforcement learning (RL) in aspects such as multi-task learning, sample efficiency, and high-level task planning. In this survey, we provide a comprehensive review of the existing literature in LLM-enhanced RL and summarize its characteristics compared to conventional RL methods, aiming to clarify the research scope and directions for future studies. Utilizing the classical agent-environment interaction paradigm, we propose a structured taxonomy to systematically categorize LLMs' functionalities in RL, including four roles: information processor, reward designer, decision-maker, and generator. For each role, we summarize the methodologies, analyze the specific RL challenges that are mitigated, and provide insights into future directions. Lastly, a comparative analysis of each role, potential applications, prospective opportunities, and challenges of the LLM-enhanced RL are discussed. By proposing this taxonomy, we aim to provide a framework for researchers to effectively leverage LLMs in the RL field, potentially accelerating RL applications in complex applications such as robotics, autonomous driving, and energy systems. △ Less

Submitted 29 October, 2024; v1 submitted 30 March, 2024; originally announced April 2024.

Comments: 22 pages (including bibliography), 6 figures

GOMA: Proactive Embodied Cooperative Communication via Goal-Oriented Mental Alignment

Authors: Lance Ying, Kunal Jha, Shivam Aarya, Joshua B. Tenenbaum, Antonio Torralba, Tianmin Shu

Abstract: Verbal communication plays a crucial role in human cooperation, particularly when the partners only have incomplete information about the task, environment, and each other's mental state. In this paper, we propose a novel cooperative communication framework, Goal-Oriented Mental Alignment (GOMA). GOMA formulates verbal communication as a planning problem that minimizes the misalignment between the… ▽ More Verbal communication plays a crucial role in human cooperation, particularly when the partners only have incomplete information about the task, environment, and each other's mental state. In this paper, we propose a novel cooperative communication framework, Goal-Oriented Mental Alignment (GOMA). GOMA formulates verbal communication as a planning problem that minimizes the misalignment between the parts of agents' mental states that are relevant to the goals. This approach enables an embodied assistant to reason about when and how to proactively initialize communication with humans verbally using natural language to help achieve better cooperation. We evaluate our approach against strong baselines in two challenging environments, Overcooked (a multiplayer game) and VirtualHome (a household simulator). Our experimental results demonstrate that large language models struggle with generating meaningful communication that is grounded in the social and physical context. In contrast, our approach can successfully generate concise verbal communication for the embodied assistant to effectively boost the performance of the cooperation as well as human users' perception of the assistant. △ Less

Submitted 14 January, 2025; v1 submitted 16 March, 2024; originally announced March 2024.

Comments: 8 pages, 5 figures

MMToM-QA: Multimodal Theory of Mind Question Answering

Authors: Chuanyang Jin, Yutong Wu, Jing Cao, Jiannan Xiang, Yen-Ling Kuo, Zhiting Hu, Tomer Ullman, Antonio Torralba, Joshua B. Tenenbaum, Tianmin Shu

Abstract: Theory of Mind (ToM), the ability to understand people's mental states, is an essential ingredient for developing machines with human-level social intelligence. Recent machine learning models, particularly large language models, seem to show some aspects of ToM understanding. However, existing ToM benchmarks use unimodal datasets - either video or text. Human ToM, on the other hand, is more than v… ▽ More Theory of Mind (ToM), the ability to understand people's mental states, is an essential ingredient for developing machines with human-level social intelligence. Recent machine learning models, particularly large language models, seem to show some aspects of ToM understanding. However, existing ToM benchmarks use unimodal datasets - either video or text. Human ToM, on the other hand, is more than video or text understanding. People can flexibly reason about another person's mind based on conceptual representations (e.g., goals, beliefs, plans) extracted from any available data. To address this, we introduce a multimodal Theory of Mind question answering (MMToM-QA) benchmark. MMToM-QA comprehensively evaluates machine ToM both on multimodal data and on different kinds of unimodal data about a person's activity in a household environment. To engineer multimodal ToM capacity, we propose a novel method, BIP-ALM (Bayesian Inverse Planning Accelerated by Language Models). BIP-ALM extracts unified representations from multimodal data and utilizes language models for scalable Bayesian inverse planning. We conducted a systematic comparison of human performance, BIP-ALM, and state-of-the-art models, including GPT-4. The experiments demonstrate that large language models and large multimodal models still lack robust ToM capacity. BIP-ALM, on the other hand, shows promising results, by leveraging the power of both model-based mental inference and language models. △ Less

Submitted 15 June, 2024; v1 submitted 16 January, 2024; originally announced January 2024.

Comments: ACL 2024. 26 pages, 11 figures, 7 tables

Free Lunch for Federated Remote Sensing Target Fine-Grained Classification: A Parameter-Efficient Framework

Authors: Shengchao Chen, Ting Shu, Huan Zhao, Jiahao Wang, Sufen Ren, Lina Yang

Abstract: Remote Sensing Target Fine-grained Classification (TFGC) is of great significance in both military and civilian fields. Due to location differences, growth in data size, and centralized server storage constraints, these data are usually stored under different databases across regions/countries. However, privacy laws and national security concerns constrain researchers from accessing these sensitiv… ▽ More Remote Sensing Target Fine-grained Classification (TFGC) is of great significance in both military and civilian fields. Due to location differences, growth in data size, and centralized server storage constraints, these data are usually stored under different databases across regions/countries. However, privacy laws and national security concerns constrain researchers from accessing these sensitive remote sensing images for further analysis. Additionally, low-resource remote sensing devices encounter challenges in terms of communication overhead and efficiency when dealing with the ever-increasing data and model scales. To solve the above challenges, this paper proposes a novel Privacy-Reserving TFGC Framework based on Federated Learning, dubbed PRFL. The proposed framework allows each client to learn global and local knowledge to enhance the local representation of private data in environments with extreme statistical heterogeneity (non. Independent and Identically Distributed, IID). Thus, it provides highly customized models to clients with differentiated data distributions. Moreover, the framework minimizes communication overhead and improves efficiency while ensuring satisfactory performance, thereby enhancing robustness and practical applicability under resource-scarce conditions. We demonstrate the effectiveness of the proposed PRFL on the classical TFGC task by leveraging four public datasets. △ Less

Submitted 2 January, 2024; originally announced January 2024.

Comments: Under Review, 23 pages, 3 figures, 12 tables

Language Models, Agent Models, and World Models: The LAW for Machine Reasoning and Planning

Authors: Zhiting Hu, Tianmin Shu

Abstract: Despite their tremendous success in many applications, large language models often fall short of consistent reasoning and planning in various (language, embodied, and social) scenarios, due to inherent limitations in their inference, learning, and modeling capabilities. In this position paper, we present a new perspective of machine reasoning, LAW, that connects the concepts of Language models, Ag… ▽ More Despite their tremendous success in many applications, large language models often fall short of consistent reasoning and planning in various (language, embodied, and social) scenarios, due to inherent limitations in their inference, learning, and modeling capabilities. In this position paper, we present a new perspective of machine reasoning, LAW, that connects the concepts of Language models, Agent models, and World models, for more robust and versatile reasoning capabilities. In particular, we propose that world and agent models are a better abstraction of reasoning, that introduces the crucial elements of deliberate human-like reasoning, including beliefs about the world and other agents, anticipation of consequences, goals/rewards, and strategic planning. Crucially, language models in LAW serve as a backend to implement the system or its elements and hence provide the computational power and adaptability. We review the recent studies that have made relevant progress and discuss future research directions towards operationalizing the LAW framework. △ Less

Submitted 8 December, 2023; originally announced December 2023.

Comments: Position paper. Accompanying NeurIPS2023 Tutorial: https://sites.google.com/view/neurips2023law/home

The Cultural Psychology of Large Language Models: Is ChatGPT a Holistic or Analytic Thinker?

Authors: Chuanyang Jin, Songyang Zhang, Tianmin Shu, Zhihan Cui

Abstract: The prevalent use of Large Language Models (LLMs) has necessitated studying their mental models, yielding noteworthy theoretical and practical implications. Current research has demonstrated that state-of-the-art LLMs, such as ChatGPT, exhibit certain theory of mind capabilities and possess relatively stable Big Five and/or MBTI personality traits. In addition, cognitive process features form an e… ▽ More The prevalent use of Large Language Models (LLMs) has necessitated studying their mental models, yielding noteworthy theoretical and practical implications. Current research has demonstrated that state-of-the-art LLMs, such as ChatGPT, exhibit certain theory of mind capabilities and possess relatively stable Big Five and/or MBTI personality traits. In addition, cognitive process features form an essential component of these mental models. Research in cultural psychology indicated significant differences in the cognitive processes of Eastern and Western people when processing information and making judgments. While Westerners predominantly exhibit analytical thinking that isolates things from their environment to analyze their nature independently, Easterners often showcase holistic thinking, emphasizing relationships and adopting a global viewpoint. In our research, we probed the cultural cognitive traits of ChatGPT. We employed two scales that directly measure the cognitive process: the Analysis-Holism Scale (AHS) and the Triadic Categorization Task (TCT). Additionally, we used two scales that investigate the value differences shaped by cultural thinking: the Dialectical Self Scale (DSS) and the Self-construal Scale (SCS). In cognitive process tests (AHS/TCT), ChatGPT consistently tends towards Eastern holistic thinking, but regarding value judgments (DSS/SCS), ChatGPT does not significantly lean towards the East or the West. We suggest that the result could be attributed to both the training paradigm and the training data in LLM development. We discuss the potential value of this finding for AI research and directions for future research. △ Less

Submitted 27 August, 2023; originally announced August 2023.

Neural Amortized Inference for Nested Multi-agent Reasoning

Authors: Kunal Jha, Tuan Anh Le, Chuanyang Jin, Yen-Ling Kuo, Joshua B. Tenenbaum, Tianmin Shu

Abstract: Multi-agent interactions, such as communication, teaching, and bluffing, often rely on higher-order social inference, i.e., understanding how others infer oneself. Such intricate reasoning can be effectively modeled through nested multi-agent reasoning. Nonetheless, the computational complexity escalates exponentially with each level of reasoning, posing a significant challenge. However, humans ef… ▽ More Multi-agent interactions, such as communication, teaching, and bluffing, often rely on higher-order social inference, i.e., understanding how others infer oneself. Such intricate reasoning can be effectively modeled through nested multi-agent reasoning. Nonetheless, the computational complexity escalates exponentially with each level of reasoning, posing a significant challenge. However, humans effortlessly perform complex social inferences as part of their daily lives. To bridge the gap between human-like inference capabilities and computational limitations, we propose a novel approach: leveraging neural networks to amortize high-order social inference, thereby expediting nested multi-agent reasoning. We evaluate our method in two challenging multi-agent interaction domains. The experimental results demonstrate that our method is computationally efficient while exhibiting minimal degradation in accuracy. △ Less

Submitted 21 August, 2023; originally announced August 2023.

Comments: 8 pages, 10 figures

Countering Eavesdroppers with Meta-learning-based Cooperative Ambient Backscatter Communications

Authors: Nam H. Chu, Nguyen Van Huynh, Diep N. Nguyen, Dinh Thai Hoang, Shimin Gong, Tao Shu, Eryk Dutkiewicz, Khoa T. Phan

Abstract: This article introduces a novel lightweight framework using ambient backscattering communications to counter eavesdroppers. In particular, our framework divides an original message into two parts: (i) the active-transmit message transmitted by the transmitter using conventional RF signals and (ii) the backscatter message transmitted by an ambient backscatter tag that backscatters upon the active s… ▽ More This article introduces a novel lightweight framework using ambient backscattering communications to counter eavesdroppers. In particular, our framework divides an original message into two parts: (i) the active-transmit message transmitted by the transmitter using conventional RF signals and (ii) the backscatter message transmitted by an ambient backscatter tag that backscatters upon the active signals emitted by the transmitter. Notably, the backscatter tag does not generate its own signal, making it difficult for an eavesdropper to detect the backscattered signals unless they have prior knowledge of the system. Here, we assume that without decoding/knowing the backscatter message, the eavesdropper is unable to decode the original message. Even in scenarios where the eavesdropper can capture both messages, reconstructing the original message is a complex task without understanding the intricacies of the message-splitting mechanism. A challenge in our proposed framework is to effectively decode the backscattered signals at the receiver, often accomplished using the maximum likelihood (MLK) approach. However, such a method may require a complex mathematical model together with perfect channel state information (CSI). To address this issue, we develop a novel deep meta-learning-based signal detector that can not only effectively decode the weak backscattered signals without requiring perfect CSI but also quickly adapt to a new wireless environment with very little knowledge. Simulation results show that our proposed learning approach, without requiring perfect CSI and complex mathematical model, can achieve a bit error ratio close to that of the MLK-based approach. They also clearly show the efficiency of the proposed approach in dealing with eavesdropping attacks and the lack of training data for deep learning models in practical scenarios. △ Less

Submitted 4 August, 2023; originally announced August 2023.

Diagnosis, Feedback, Adaptation: A Human-in-the-Loop Framework for Test-Time Policy Adaptation

Authors: Andi Peng, Aviv Netanyahu, Mark Ho, Tianmin Shu, Andreea Bobu, Julie Shah, Pulkit Agrawal

Abstract: Policies often fail due to distribution shift -- changes in the state and reward that occur when a policy is deployed in new environments. Data augmentation can increase robustness by making the model invariant to task-irrelevant changes in the agent's observation. However, designers don't know which concepts are irrelevant a priori, especially when different end users have different preferences a… ▽ More Policies often fail due to distribution shift -- changes in the state and reward that occur when a policy is deployed in new environments. Data augmentation can increase robustness by making the model invariant to task-irrelevant changes in the agent's observation. However, designers don't know which concepts are irrelevant a priori, especially when different end users have different preferences about how the task is performed. We propose an interactive framework to leverage feedback directly from the user to identify personalized task-irrelevant concepts. Our key idea is to generate counterfactual demonstrations that allow users to quickly identify possible task-relevant and irrelevant concepts. The knowledge of task-irrelevant concepts is then used to perform data augmentation and thus obtain a policy adapted to personalized user objectives. We present experiments validating our framework on discrete and continuous control tasks with real human users. Our method (1) enables users to better understand agent failure, (2) reduces the number of demonstrations required for fine-tuning, and (3) aligns the agent to individual user task preferences. △ Less

Submitted 13 July, 2023; v1 submitted 12 July, 2023; originally announced July 2023.

Comments: International Conference on Machine Learning (ICML) 2023

Building Cooperative Embodied Agents Modularly with Large Language Models

Authors: Hongxin Zhang, Weihua Du, Jiaming Shan, Qinhong Zhou, Yilun Du, Joshua B. Tenenbaum, Tianmin Shu, Chuang Gan

Abstract: In this work, we address challenging multi-agent cooperation problems with decentralized control, raw sensory observations, costly communication, and multi-objective tasks instantiated in various embodied environments. While previous research either presupposes a cost-free communication channel or relies on a centralized controller with shared observations, we harness the commonsense knowledge, re… ▽ More In this work, we address challenging multi-agent cooperation problems with decentralized control, raw sensory observations, costly communication, and multi-objective tasks instantiated in various embodied environments. While previous research either presupposes a cost-free communication channel or relies on a centralized controller with shared observations, we harness the commonsense knowledge, reasoning ability, language comprehension, and text generation prowess of LLMs and seamlessly incorporate them into a cognitive-inspired modular framework that integrates with perception, memory, and execution. Thus building a Cooperative Embodied Language Agent CoELA, who can plan, communicate, and cooperate with others to accomplish long-horizon tasks efficiently. Our experiments on C-WAH and TDW-MAT demonstrate that CoELA driven by GPT-4 can surpass strong planning-based methods and exhibit emergent effective communication. Though current Open LMs like LLAMA-2 still underperform, we fine-tune a CoELA with data collected with our agents and show how they can achieve promising performance. We also conducted a user study for human-agent interaction and discovered that CoELA communicating in natural language can earn more trust and cooperate more effectively with humans. Our research underscores the potential of LLMs for future research in multi-agent cooperation. Videos can be found on the project website https://vis-www.cs.umass.edu/Co-LLM-Agents/. △ Less

Submitted 17 February, 2024; v1 submitted 5 July, 2023; originally announced July 2023.

Comments: ICLR24. The first two authors contributed equally

Language Models Meet World Models: Embodied Experiences Enhance Language Models

Authors: Jiannan Xiang, Tianhua Tao, Yi Gu, Tianmin Shu, Zirui Wang, Zichao Yang, Zhiting Hu

Abstract: While large language models (LMs) have shown remarkable capabilities across numerous tasks, they often struggle with simple reasoning and planning in physical environments, such as understanding object permanence or planning household activities. The limitation arises from the fact that LMs are trained only on written text and miss essential embodied knowledge and skills. In this paper, we propose… ▽ More While large language models (LMs) have shown remarkable capabilities across numerous tasks, they often struggle with simple reasoning and planning in physical environments, such as understanding object permanence or planning household activities. The limitation arises from the fact that LMs are trained only on written text and miss essential embodied knowledge and skills. In this paper, we propose a new paradigm of enhancing LMs by finetuning them with world models, to gain diverse embodied knowledge while retaining their general language capabilities. Our approach deploys an embodied agent in a world model, particularly a simulator of the physical world (VirtualHome), and acquires a diverse set of embodied experiences through both goal-oriented planning and random exploration. These experiences are then used to finetune LMs to teach diverse abilities of reasoning and acting in the physical world, e.g., planning and completing goals, object permanence and tracking, etc. Moreover, it is desirable to preserve the generality of LMs during finetuning, which facilitates generalizing the embodied knowledge across tasks rather than being tied to specific simulations. We thus further introduce the classical (EWC) for selective weight updates, combined with low-rank adapters (LoRA) for training efficiency. Extensive experiments show our approach substantially improves base LMs on 18 downstream tasks by 64.28% on average. In particular, the small LMs (1.3B, 6B, and 13B) enhanced by our approach match or even outperform much larger LMs (e.g., ChatGPT). △ Less

Submitted 28 October, 2023; v1 submitted 17 May, 2023; originally announced May 2023.

MASK-CNN-Transformer For Real-Time Multi-Label Weather Recognition

Authors: Shengchao Chen, Ting Shu, Huan Zhao, Yuan Yan Tang

Abstract: Weather recognition is an essential support for many practical life applications, including traffic safety, environment, and meteorology. However, many existing related works cannot comprehensively describe weather conditions due to their complex co-occurrence dependencies. This paper proposes a novel multi-label weather recognition model considering these dependencies. The proposed model called M… ▽ More Weather recognition is an essential support for many practical life applications, including traffic safety, environment, and meteorology. However, many existing related works cannot comprehensively describe weather conditions due to their complex co-occurrence dependencies. This paper proposes a novel multi-label weather recognition model considering these dependencies. The proposed model called MASK-Convolutional Neural Network-Transformer (MASK-CT) is based on the Transformer, the convolutional process, and the MASK mechanism. The model employs multiple convolutional layers to extract features from weather images and a Transformer encoder to calculate the probability of each weather condition based on the extracted features. To improve the generalization ability of MASK-CT, a MASK mechanism is used during the training phase. The effect of the MASK mechanism is explored and discussed. The Mask mechanism randomly withholds some information from one-pair training instances (one image and its corresponding label). There are two types of MASK methods. Specifically, MASK-I is designed and deployed on the image before feeding it into the weather feature extractor and MASK-II is applied to the image label. The Transformer encoder is then utilized on the randomly masked image features and labels. The experimental results from various real-world weather recognition datasets demonstrate that the proposed MASK-CT model outperforms state-of-the-art methods. Furthermore, the high-speed dynamic real-time weather recognition capability of the MASK-CT is evaluated. △ Less

Submitted 19 August, 2023; v1 submitted 28 April, 2023; originally announced April 2023.

Comments: Have been accepted. Appears in Knowledge-Based Systems, see https://www.sciencedirect.com/science/article/pii/S0950705123006317

Journal ref: Knowledge-Based Systems, 110881 (2023)

TempEE: Temporal-Spatial Parallel Transformer for Radar Echo Extrapolation Beyond Auto-Regression

Authors: Shengchao Chen, Ting Shu, Huan Zhao, Guo Zhong, Xunlai Chen

Abstract: Meteorological radar reflectivity data (i.e. radar echo) significantly influences precipitation prediction. It can facilitate accurate and expeditious forecasting of short-term heavy rainfall bypassing the need for complex Numerical Weather Prediction (NWP) models. In comparison to conventional models, Deep Learning (DL)-based radar echo extrapolation algorithms exhibit higher effectiveness and ef… ▽ More Meteorological radar reflectivity data (i.e. radar echo) significantly influences precipitation prediction. It can facilitate accurate and expeditious forecasting of short-term heavy rainfall bypassing the need for complex Numerical Weather Prediction (NWP) models. In comparison to conventional models, Deep Learning (DL)-based radar echo extrapolation algorithms exhibit higher effectiveness and efficiency. Nevertheless, the development of reliable and generalized echo extrapolation algorithm is impeded by three primary challenges: cumulative error spreading, imprecise representation of sparsely distributed echoes, and inaccurate description of non-stationary motion processes. To tackle these challenges, this paper proposes a novel radar echo extrapolation algorithm called Temporal-Spatial Parallel Transformer, referred to as TempEE. TempEE avoids using auto-regression and instead employs a one-step forward strategy to prevent cumulative error spreading during the extrapolation process. Additionally, we propose the incorporation of a Multi-level Temporal-Spatial Attention mechanism to improve the algorithm's capability of capturing both global and local information while emphasizing task-related regions, including sparse echo representations, in an efficient manner. Furthermore, the algorithm extracts spatio-temporal representations from continuous echo images using a parallel encoder to model the non-stationary motion process for echo extrapolation. The superiority of our TempEE has been demonstrated in the context of the classic radar echo extrapolation task, utilizing a real-world dataset. Extensive experiments have further validated the efficacy and indispensability of various components within TempEE. △ Less

Submitted 14 September, 2023; v1 submitted 27 April, 2023; originally announced April 2023.

Comments: Have been accepted by IEEE Transactions on Geoscience and Remote Sensing, see https://ieeexplore.ieee.org/document/10238744

Journal ref: IEEE Transactions on Geoscience and Remote Sensing 61, 5108914 (2023)

Mutilmodal Feature Extraction and Attention-based Fusion for Emotion Estimation in Videos

Authors: Tao Shu, Xinke Wang, Ruotong Wang, Chuang Chen, Yixin Zhang, Xiao Sun

Abstract: The continuous improvement of human-computer interaction technology makes it possible to compute emotions. In this paper, we introduce our submission to the CVPR 2023 Competition on Affective Behavior Analysis in-the-wild (ABAW). Sentiment analysis in human-computer interaction should, as far as possible Start with multiple dimensions, fill in the single imperfect emotion channel, and finally dete… ▽ More The continuous improvement of human-computer interaction technology makes it possible to compute emotions. In this paper, we introduce our submission to the CVPR 2023 Competition on Affective Behavior Analysis in-the-wild (ABAW). Sentiment analysis in human-computer interaction should, as far as possible Start with multiple dimensions, fill in the single imperfect emotion channel, and finally determine the emotion tendency by fitting multiple results. Therefore, We exploited multimodal features extracted from video of different lengths from the competition dataset, including audio, pose and images. Well-informed emotion representations drive us to propose a Attention-based multimodal framework for emotion estimation. Our system achieves the performance of 0.361 on the validation dataset. The code is available at [https://github.com/xkwangcn/ABAW-5th-RT-IAI]. △ Less

Submitted 18 March, 2023; originally announced March 2023.

Comments: 5 pages, 1 figures

Dynamic Resource Allocation for Metaverse Applications with Deep Reinforcement Learning

Authors: Nam H. Chu, Diep N. Nguyen, Dinh Thai Hoang, Khoa T. Phan, Eryk Dutkiewicz, Dusit Niyato, Tao Shu

Abstract: This work proposes a novel framework to dynamically and effectively manage and allocate different types of resources for Metaverse applications, which are forecasted to demand massive resources of various types that have never been seen before. Specifically, by studying functions of Metaverse applications, we first propose an effective solution to divide applications into groups, namely MetaInstan… ▽ More This work proposes a novel framework to dynamically and effectively manage and allocate different types of resources for Metaverse applications, which are forecasted to demand massive resources of various types that have never been seen before. Specifically, by studying functions of Metaverse applications, we first propose an effective solution to divide applications into groups, namely MetaInstances, where common functions can be shared among applications to enhance resource usage efficiency. Then, to capture the real-time, dynamic, and uncertain characteristics of request arrival and application departure processes, we develop a semi-Markov decision process-based framework and propose an intelligent algorithm that can gradually learn the optimal admission policy to maximize the revenue and resource usage efficiency for the Metaverse service provider and at the same time enhance the Quality-of-Service for Metaverse users. Extensive simulation results show that our proposed approach can achieve up to 120% greater revenue for the Metaverse service providers and up to 178.9% higher acceptance probability for Metaverse application requests than those of other baselines. △ Less

Submitted 26 February, 2023; originally announced February 2023.

Comments: To be published in the Proceedings of the IEEE WCNC 2023

NOPA: Neurally-guided Online Probabilistic Assistance for Building Socially Intelligent Home Assistants

Authors: Xavier Puig, Tianmin Shu, Joshua B. Tenenbaum, Antonio Torralba

Abstract: In this work, we study how to build socially intelligent robots to assist people in their homes. In particular, we focus on assistance with online goal inference, where robots must simultaneously infer humans' goals and how to help them achieve those goals. Prior assistance methods either lack the adaptivity to adjust helping strategies (i.e., when and how to help) in response to uncertainty about… ▽ More In this work, we study how to build socially intelligent robots to assist people in their homes. In particular, we focus on assistance with online goal inference, where robots must simultaneously infer humans' goals and how to help them achieve those goals. Prior assistance methods either lack the adaptivity to adjust helping strategies (i.e., when and how to help) in response to uncertainty about goals or the scalability to conduct fast inference in a large goal space. Our NOPA (Neurally-guided Online Probabilistic Assistance) method addresses both of these challenges. NOPA consists of (1) an online goal inference module combining neural goal proposals with inverse planning and particle filtering for robust inference under uncertainty, and (2) a helping planner that discovers valuable subgoals to help with and is aware of the uncertainty in goal inference. We compare NOPA against multiple baselines in a new embodied AI assistance challenge: Online Watch-And-Help, in which a helper agent needs to simultaneously watch a main agent's action, infer its goal, and help perform a common household task faster in realistic virtual home environments. Experiments show that our helper agent robustly updates its goal inference and adapts its helping plans to the changing level of uncertainty. △ Less

Submitted 12 January, 2023; originally announced January 2023.

Comments: Project website: https://www.tshu.io/online_watch_and_help. Code: https://github.com/xavierpuigf/online_watch_and_help

State Space Closure: Revisiting Endless Online Level Generation via Reinforcement Learning

Authors: Ziqi Wang, Tianye Shu, Jialin Liu

Abstract: In this paper, we revisit endless online level generation with the recently proposed experience-driven procedural content generation via reinforcement learning (EDRL) framework. Inspired by an observation that EDRL tends to generate recurrent patterns, we formulate a notion of state space closure which makes any stochastic state appeared possibly in an infinite-horizon online generation process ca… ▽ More In this paper, we revisit endless online level generation with the recently proposed experience-driven procedural content generation via reinforcement learning (EDRL) framework. Inspired by an observation that EDRL tends to generate recurrent patterns, we formulate a notion of state space closure which makes any stochastic state appeared possibly in an infinite-horizon online generation process can be found within a finite-horizon. Through theoretical analysis, we find that even though state space closure arises a concern about diversity, it generalises EDRL trained with a finite-horizon to the infinite-horizon scenario without deterioration of content quality. Moreover, we verify the quality and the diversity of contents generated by EDRL via empirical studies, on the widely used Super Mario Bros. benchmark. Experimental results reveal that the diversity of levels generated by EDRL is limited due to the state space closure, whereas their quality does not deteriorate in a horizon which is longer than the one specified in the training. Concluding our outcomes and analysis, future work on endless online level generation via reinforcement learning should address the issue of diversity while assuring the occurrence of state space closure and quality. △ Less

Submitted 24 March, 2023; v1 submitted 6 December, 2022; originally announced December 2022.

Comments: Accepted by the IEEE Transactions on Games

ACM Class: I.2.6

Discovering Generalizable Spatial Goal Representations via Graph-based Active Reward Learning

Authors: Aviv Netanyahu, Tianmin Shu, Joshua Tenenbaum, Pulkit Agrawal

Abstract: In this work, we consider one-shot imitation learning for object rearrangement tasks, where an AI agent needs to watch a single expert demonstration and learn to perform the same task in different environments. To achieve a strong generalization, the AI agent must infer the spatial goal specification for the task. However, there can be multiple goal specifications that fit the given demonstration.… ▽ More In this work, we consider one-shot imitation learning for object rearrangement tasks, where an AI agent needs to watch a single expert demonstration and learn to perform the same task in different environments. To achieve a strong generalization, the AI agent must infer the spatial goal specification for the task. However, there can be multiple goal specifications that fit the given demonstration. To address this, we propose a reward learning approach, Graph-based Equivalence Mappings (GEM), that can discover spatial goal representations that are aligned with the intended goal specification, enabling successful generalization in unseen environments. Specifically, GEM represents a spatial goal specification by a reward function conditioned on i) a graph indicating important spatial relationships between objects and ii) state equivalence mappings for each edge in the graph indicating invariant properties of the corresponding relationship. GEM combines inverse reinforcement learning and active reward learning to efficiently improve the reward function by utilizing the graph structure and domain randomization enabled by the equivalence mappings. We conducted experiments with simulated oracles and with human subjects. The results show that GEM can drastically improve the generalizability of the learned goal representations over strong baselines. △ Less

Submitted 24 November, 2022; originally announced November 2022.

Comments: ICML 2022, the first two authors contributed equally, project page https://www.tshu.io/GEM

Stateful active facilitator: Coordination and Environmental Heterogeneity in Cooperative Multi-Agent Reinforcement Learning

Authors: Dianbo Liu, Vedant Shah, Oussama Boussif, Cristian Meo, Anirudh Goyal, Tianmin Shu, Michael Mozer, Nicolas Heess, Yoshua Bengio

Abstract: In cooperative multi-agent reinforcement learning, a team of agents works together to achieve a common goal. Different environments or tasks may require varying degrees of coordination among agents in order to achieve the goal in an optimal way. The nature of coordination will depend on the properties of the environment -- its spatial layout, distribution of obstacles, dynamics, etc. We term this… ▽ More In cooperative multi-agent reinforcement learning, a team of agents works together to achieve a common goal. Different environments or tasks may require varying degrees of coordination among agents in order to achieve the goal in an optimal way. The nature of coordination will depend on the properties of the environment -- its spatial layout, distribution of obstacles, dynamics, etc. We term this variation of properties within an environment as heterogeneity. Existing literature has not sufficiently addressed the fact that different environments may have different levels of heterogeneity. We formalize the notions of coordination level and heterogeneity level of an environment and present HECOGrid, a suite of multi-agent RL environments that facilitates empirical evaluation of different MARL approaches across different levels of coordination and environmental heterogeneity by providing a quantitative control over coordination and heterogeneity levels of the environment. Further, we propose a Centralized Training Decentralized Execution learning approach called Stateful Active Facilitator (SAF) that enables agents to work efficiently in high-coordination and high-heterogeneity environments through a differentiable and shared knowledge source used during training and dynamic selection from a shared pool of policies. We evaluate SAF and compare its performance against baselines IPPO and MAPPO on HECOGrid. Our results show that SAF consistently outperforms the baselines across different tasks and different heterogeneity and coordination levels. We release the code for HECOGrid as well as all our experiments. △ Less

Submitted 6 October, 2023; v1 submitted 4 October, 2022; originally announced October 2022.

Comments: Published at ICLR 2023

PTSG: a test generation tool based on extended finite state machine

Authors: Zhijie Pan, Ting Shu, Zuohua Ding

Abstract: The Extended Finite State Machine (EFSM) is one of the most popular modeling approaches for model-based testing. However, EFSM-based test case generation is susceptible to the infeasible (inexecutable) path problem, which stems from the conflict of predicates (guards) between transitions in the path. Therefore, in order to derive feasible test cases, a test generation algorithm needs to dynamicall… ▽ More The Extended Finite State Machine (EFSM) is one of the most popular modeling approaches for model-based testing. However, EFSM-based test case generation is susceptible to the infeasible (inexecutable) path problem, which stems from the conflict of predicates (guards) between transitions in the path. Therefore, in order to derive feasible test cases, a test generation algorithm needs to dynamically acquire information about the model and verify the feasibility of the generated test path through the simulation execution of the model. The traditional method of constructing executable models using hard-coding for different EFSM models under test has limitations such as inflexibility, time-consuming and error-prone. To address this issue, this paper develops an open source test generation tool for testing EFSM-specified systems, PTSG, to support the automatic generation of executable test cases. It decouples the EFSM model description, parsing and simulation execution functions from the test generation algorithm, which can effectively improve the efficiency and quality of test generation. In particular, PTSG first uses a well-designed JSON syntax to describe the specific EFSM under test. Then, based on the model description file, it uses lexical and syntactic parsers to dynamically extract model information to construct various model objects in memory such as state configurations, transitions, etc. Finally, the tool provide a series of service interfaces to support model information acquisition, transition feasibility evaluation, and model simulation execution. A case study of test sequence generation for the SCP protocol model demonstrates the capability and promise of the PTSG to serve executable test cases. △ Less

Submitted 21 September, 2022; originally announced September 2022.

Effects of Archive Size on Computation Time and Solution Quality for Multi-Objective Optimization

Authors: Tianye Shu, Ke Shang, Hisao Ishibuchi, Yang Nan

Abstract: An unbounded external archive has been used to store all nondominated solutions found by an evolutionary multi-objective optimization algorithm in some studies. It has been shown that a selected solution subset from the stored solutions is often better than the final population. However, the use of the unbounded archive is not always realistic. When the number of examined solutions is huge, we mus… ▽ More An unbounded external archive has been used to store all nondominated solutions found by an evolutionary multi-objective optimization algorithm in some studies. It has been shown that a selected solution subset from the stored solutions is often better than the final population. However, the use of the unbounded archive is not always realistic. When the number of examined solutions is huge, we must pre-specify the archive size. In this study, we examine the effects of the archive size on three aspects: (i) the quality of the selected final solution set, (ii) the total computation time for the archive maintenance and the final solution set selection, and (iii) the required memory size. Unsurprisingly, the increase of the archive size improves the final solution set quality. Interestingly, the total computation time of a medium-size archive is much larger than that of a small-size archive and a huge-size archive (e.g., an unbounded archive). To decrease the computation time, we examine two ideas: periodical archive update and archiving only in later generations. Compared with updating the archive at every generation, the first idea can obtain almost the same final solution set quality using a much shorter computation time at the cost of a slight increase of the memory size. The second idea drastically decreases the computation time at the cost of a slight deterioration of the final solution set quality. Based on our experimental results, some suggestions are given about how to appropriately choose an archiving strategy and an archive size. △ Less

Submitted 2 November, 2022; v1 submitted 7 September, 2022; originally announced September 2022.

Technical Report: Assisting Backdoor Federated Learning with Whole Population Knowledge Alignment

Authors: Tian Liu, Xueyang Hu, Tao Shu

Abstract: Due to the distributed nature of Federated Learning (FL), researchers have uncovered that FL is vulnerable to backdoor attacks, which aim at injecting a sub-task into the FL without corrupting the performance of the main task. Single-shot backdoor attack achieves high accuracy on both the main task and backdoor sub-task when injected at the FL model convergence. However, the early-injected single-… ▽ More Due to the distributed nature of Federated Learning (FL), researchers have uncovered that FL is vulnerable to backdoor attacks, which aim at injecting a sub-task into the FL without corrupting the performance of the main task. Single-shot backdoor attack achieves high accuracy on both the main task and backdoor sub-task when injected at the FL model convergence. However, the early-injected single-shot backdoor attack is ineffective because: (1) the maximum backdoor effectiveness is not reached at injection because of the dilution effect from normal local updates; (2) the backdoor effect decreases quickly as the backdoor will be overwritten by the newcoming normal local updates. In this paper, we strengthen the early-injected single-shot backdoor attack utilizing FL model information leakage. We show that the FL convergence can be expedited if the client trains on a dataset that mimics the distribution and gradients of the whole population. Based on this observation, we proposed a two-phase backdoor attack, which includes a preliminary phase for the subsequent backdoor attack. In the preliminary phase, the attacker-controlled client first launches a whole population distribution inference attack and then trains on a locally crafted dataset that is aligned with both the gradient and inferred distribution. Benefiting from the preliminary phase, the later injected backdoor achieves better effectiveness as the backdoor effect will be less likely to be diluted by the normal model updates. Extensive experiments are conducted on MNIST dataset under various data heterogeneity settings to evaluate the effectiveness of the proposed backdoor attack. Results show that the proposed backdoor outperforms existing backdoor attacks in both success rate and longevity, even when defense mechanisms are in place. △ Less

Submitted 25 July, 2022; originally announced July 2022.

RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learning

Authors: Mingkai Deng, Jianyu Wang, Cheng-Ping Hsieh, Yihan Wang, Han Guo, Tianmin Shu, Meng Song, Eric P. Xing, Zhiting Hu

Abstract: Prompting has shown impressive success in enabling large pretrained language models (LMs) to perform diverse NLP tasks, especially when only few downstream data are available. Automatically finding the optimal prompt for each task, however, is challenging. Most existing work resorts to tuning soft prompt (e.g., embeddings) which falls short of interpretability, reusability across LMs, and applicab… ▽ More Prompting has shown impressive success in enabling large pretrained language models (LMs) to perform diverse NLP tasks, especially when only few downstream data are available. Automatically finding the optimal prompt for each task, however, is challenging. Most existing work resorts to tuning soft prompt (e.g., embeddings) which falls short of interpretability, reusability across LMs, and applicability when gradients are not accessible. Discrete prompt, on the other hand, is difficult to optimize, and is often created by "enumeration (e.g., paraphrasing)-then-selection" heuristics that do not explore the prompt space systematically. This paper proposes RLPrompt, an efficient discrete prompt optimization approach with reinforcement learning (RL). RLPrompt formulates a parameter-efficient policy network that generates the desired discrete prompt after training with reward. To overcome the complexity and stochasticity of reward signals by the large LM environment, we incorporate effective reward stabilization that substantially enhances the training efficiency. RLPrompt is flexibly applicable to different types of LMs, such as masked (e.g., BERT) and left-to-right models (e.g., GPTs), for both classification and generation tasks. Experiments on few-shot classification and unsupervised text style transfer show superior performance over a wide range of existing finetuning or prompting methods. Interestingly, the resulting optimized prompts are often ungrammatical gibberish text; and surprisingly, those gibberish prompts are transferrable between different LMs to retain significant performance, indicating LM prompting may not follow human language patterns. △ Less

Submitted 22 October, 2022; v1 submitted 25 May, 2022; originally announced May 2022.

Comments: EMNLP 2022 Camera Ready. Code available at https://github.com/mingkaid/rl-prompt

MetaSlicing: A Novel Resource Allocation Framework for Metaverse

Authors: Nam H. Chu, Dinh Thai Hoang, Diep N. Nguyen, Khoa T. Phan, Eryk Dutkiewicz, Dusit Niyato, Tao Shu

Abstract: Creating and maintaining the Metaverse requires enormous resources that have never been seen before, especially computing resources for intensive data processing to support the Extended Reality, enormous storage resources, and massive networking resources for maintaining ultra high-speed and low-latency connections. Therefore, this work aims to propose a novel framework, namely MetaSlicing, that c… ▽ More Creating and maintaining the Metaverse requires enormous resources that have never been seen before, especially computing resources for intensive data processing to support the Extended Reality, enormous storage resources, and massive networking resources for maintaining ultra high-speed and low-latency connections. Therefore, this work aims to propose a novel framework, namely MetaSlicing, that can provide a highly effective and comprehensive solution in managing and allocating different types of resources for Metaverse applications. In particular, by observing that Metaverse applications may have common functions, we first propose grouping applications into clusters, called MetaInstances. In a MetaInstance, common functions can be shared among applications. As such, the same resources can be used by multiple applications simultaneously, thereby enhancing resource utilization dramatically.To address the real-time characteristic and resource demand's dynamic and uncertainty in the Metaverse, we develop an effective framework based on the semi-Markov decision process and propose an intelligent admission control algorithm that can maximize resource utilization and enhance the Quality-of-Service for end-users. Extensive simulation results show that our proposed solution outperforms the Greedy-based policies by up to 80% and 47% in terms of long-term revenue for Metaverse providers and request acceptance probability, respectively. △ Less

Submitted 26 February, 2023; v1 submitted 23 May, 2022; originally announced May 2022.

Comments: Revised figures, fix typos