arXiv subjects

Jinghe Wang

Publications and source records attributed to Jinghe Wang.

11 recordsLinked to original sources

arXiv2026

GeniWorld: A Generalizable Interactive World Model for Robotic Manipulation via Visual Actions

Generalist robot policies exhibit strong capabilities, but their robustness in complex and unseen environments remains limited. Scaling robot learning and evaluation in diverse real-world environments remains costly and challenging. Action-conditioned world models offer a promising alternative, but they often suffer from limited action controllability and poor generalization to out-of-distribution (OOD) scenarios. To this end, we present GeniWorld, an interactive world model for robots that generalizes robustly across unseen scenarios. Building on pretrained video generative models, we use URDF-based rendering to transform numerical actions into visual action representations, enabling spatially grounded action control. By explicitly decoupling embodiment kinematics from environmental dynamics, our model mitigates scene overfitting and facilitates modeling of robot-environment interactions. To achieve closed-loop control, we construct an autoregressive video prediction model integrated with high-frequency robot kinematic control, enabling interaction with both robot policies and human teleoperators. In our experiments, even when trained solely on limited fixed-scene data, our model achieves superior in-domain performance and robust zero-shot generalization to highly randomized, unseen environments. For downstream applications, GeniWorld serves as a scalable policy evaluator that remains reliable under environmental perturbations. Furthermore, even with limited real-world demonstrations, GeniWorld generates diverse manipulation trajectories within the world model, improving downstream policy performance and robustness in complex environments.

cs.RO↗

arXiv2026

Learning with Challenges: Adaptive Difficulty-Aware Data Generation for Mobile GUI Agent Training

Large-scale, high-quality interaction trajectories are essential for advancing mobile Graphical User Interface (GUI) agents. While existing methods typically rely on labor-intensive human demonstrations or automated model exploration to generate GUI trajectories, they lack fine-grained control over task difficulty. This fundamentally restricts learning effectiveness due to the mismatch between the training difficulty and the agent's capabilities. Inspired by how humans acquire skills through progressively challenging tasks, we propose MobileGen, a novel data generation framework that adaptively aligns training difficulty with the GUI agent's capability frontier. Specifically, MobileGen explicitly decouples task difficulty into structural (e.g., trajectory length) and semantic (e.g., task goal) dimensions. It then iteratively evaluates the agent on a curated prior dataset to construct a systematic profile of its capability frontier across these two dimensions. With this profile, the probability distribution of task difficulty is adaptively computed, from which the target difficulty for the next round of training can be sampled. Guided by the sampled difficulty, a multi-agent controllable generator is finally used to synthesize high-quality interaction trajectories along with corresponding task instructions. Extensive experiments show that MobileGen consistently outperforms existing data generation methods by improving the average performance of GUI agents by 1.57 times across multiple challenging benchmarks. This highlights the importance of capability-aligned data generation for effective mobile GUI agent training.

cs.AI↗

arXiv2025

IGen: Scalable Data Generation for Robot Learning from Open-World Images

The rise of generalist robotic policies has created an exponential demand for large-scale training data. However, on-robot data collection is labor-intensive and often limited to specific environments. In contrast, open-world images capture a vast diversity of real-world scenes that naturally align with robotic manipulation tasks, offering a promising avenue for low-cost, large-scale robot data acquisition. Despite this potential, the lack of associated robot actions hinders the practical use of open-world images for robot learning, leaving this rich visual resource largely unexploited. To bridge this gap, we propose IGen, a framework that scalably generates realistic visual observations and executable actions from open-world images. IGen first converts unstructured 2D pixels into structured 3D scene representations suitable for scene understanding and manipulation. It then leverages the reasoning capabilities of vision-language models to transform scene-specific task instructions into high-level plans and generate low-level actions as SE(3) end-effector pose sequences. From these poses, it synthesizes dynamic scene evolution and renders temporally coherent visual observations. Experiments validate the high quality of visuomotor data generated by IGen, and show that policies trained solely on IGen-synthesized data achieve performance comparable to those trained on real-world data. This highlights the potential of IGen to support scalable data generation from open-world images for generalist robotic policy training.

cs.RO↗

arXiv2025

Collaborative Belief Reasoning with LLMs for Efficient Multi-Agent Collaboration

Effective real-world multi-agent collaboration requires not only accurate planning but also the ability to reason about collaborators' intents--a crucial capability for avoiding miscoordination and redundant communication under partial observable environments. Due to their strong planning and reasoning capabilities, large language models (LLMs) have emerged as promising autonomous agents for collaborative task solving. However, existing collaboration frameworks for LLMs overlook their reasoning potential for dynamic intent inference, and thus produce inconsistent plans and redundant communication, reducing collaboration efficiency. To bridge this gap, we propose CoBel-World, a novel framework that equips LLM agents with a Collaborative Belief World--an internal representation jointly modeling the physical environment and collaborators' mental states. CoBel-World enables agents to parse external open-world knowledge into structured beliefs via a symbolic belief representation module, and perform zero-shot Bayesian-style belief updates through LLM reasoning. This allows agents to proactively detect potential miscoordination (e.g., conflicting plans) and communicate adaptively. Evaluated on challenging embodied benchmarks (i.e., TDW-MAT and C-WAH), CoBel-World significantly reduces communication costs by 64-79% and improves task completion efficiency by 4-28% compared to the strongest baseline. Our results show that explicit, intent-aware belief modeling is essential for efficient and human-like collaboration in LLM-based multi-agent systems.

cs.AI↗

arXiv2024

CATP-LLM: Empowering Large Language Models for Cost-Aware Tool Planning

Utilizing large language models (LLMs) for tool planning has emerged as a promising avenue for developing general AI systems, where LLMs automatically schedule external tools (e.g., vision models) to tackle complex tasks based on task descriptions. To push this paradigm toward practical applications, it is crucial for LLMs to consider tool execution costs (e.g., execution time) for tool planning. Unfortunately, prior studies overlook the tool execution costs, leading to the generation of expensive plans whose costs outweigh their benefits in terms of task performance. To fill this gap, we propose the Cost-Aware Tool Planning with LLMs (CATP-LLM) framework, which for the first time provides a coherent design to empower LLMs for cost-aware tool planning. Specifically, To facilitate efficient concurrent tool execution and cost reduction, we design a tool planning language to enhance the LLM for creating multi-branch non-sequential plans. Moreover, we propose a cost-aware offline reinforcement learning algorithm to fine-tune the LLM to optimize the performance-cost trade-off in tool planning. In the lack of public cost-related datasets, we further present OpenCATP, the first dataset for cost-aware planning, which comprises 11,100 evaluation samples from diverse tasks. Extensive experiments show that CATP-LLM outperforms GPT-4 even when using Llama2-7B as its backbone, with the average improvement of 1.5%-93.9% in terms of plan quality. Codes and dataset are available at: https://github.com/duowuyms/OpenCATP-LLM.

cs.AI↗

arXiv2024

Rotatable Block-Controlled RIS: Bridging the Performance Gap to Element-Controlled Systems

The passive reconfigurable intelligent surface (RIS) requires numerous elements to achieve adequate array gain, which linearly increases power consumption (PC) with the number of reflection phases. To address this, this letter introduces a rotatable block-controlled RIS (BC-RIS) that preserves spectral efficiency (SE) while reducing power costs. Unlike the element-controlled RIS (EC-RIS), which necessitates independent phase control for each element, the BC-RIS uses a single phase control circuit for each block, substantially lowering power requirements. In the maximum ratio transmission, by customizing specular reflection channels through the rotation of blocks and coherently superimposing signals with optimized reflection phase of blocks, the BC-RIS achieves the same averaged SE as the EC-RIS. To counteract the added power demands from rotation, influenced by block size, we have developed a segmentation scheme to minimize overall PC. Furthermore, constraints for rotation power-related parameters have been established to enhance the energy efficiency of the BC-RIS compared to the EC-RIS. Numerical results confirm that this approach significantly improves energy efficiency while maintaining performance.

cs.IT↗

arXiv2024

Reconfigurable Intelligent Surface: Power Consumption Modeling and Practical Measurement Validation

The reconfigurable intelligent surface (RIS) has received a lot of interest because of its capacity to reconfigure the wireless communication environment in a cost- and energy-efficient way. However, the realistic power consumption modeling and measurement validation of RIS has received far too little attention. Therefore, in this work, we model the power consumption of RIS and conduct measurement validations using various RISs to fill this vacancy. Firstly, we propose a practical power consumption model of RIS. The RIS hardware is divided into three basic parts: the FPGA control board, the drive circuits, and the RIS unit cells. The power consumption of the first two parts is modeled as $P_{\text {static}}$ and that of the last part is modeled as $P_{\text {units}}$. Expressions of $P_{\text {static}}$ and $P_{\text {units}}$ vary amongst different types of RISs. Secondly, we conduct measurements on various RISs to validate the proposed model. Five different RISs including the PIN diode, varactor diode, and RF switch types are measured, and measurement results validate the generality and applicability of the proposed power consumption model of RIS. Finally, we summarize the measurement results and discuss the approaches to achieve the low-power-consumption design of RIS-assisted wireless communication systems.

cs.IT↗

arXiv2023

Static Power Consumption Modeling and Measurement of Reconfigurable Intelligent Surfaces

Reconfigurable intelligent surfaces (RISs) are anticipated to transform wireless communication in a way that is both economical and energy efficient. Revealing the practical power consumption characteristics of RISs can provide an essential toolkit for the optimal design of RIS-assisted wireless communication systems and energy efficiency performance evaluation. Based on our previous work that modeled the dynamic power consumption of RISs, we henceforth concentrate more on static power consumption. We first divide the RIS hardware into three basic parts: the FPGA control board, the drive circuits, and the RIS unit cells. The first two parts are mainly to be investigated and the last part has been modeled as the dynamic power consumption in the previous work. In this work, the power consumption of the FPGA control board is regarded as a constant value, however, that of the drive circuit is a variant that is affected by the number of control signals and its self-power consumption characteristics. Therefore, we model the power consumption of the drive circuits of various kinds of RISs, i.e., PIN diode-/Varactor diode-/RF switch-based RIS. Finally, the measurement results and typical value of static power consumption are illustrated and discussed.

cs.IT↗

arXiv2021

Smart Radio Environments

This Roadmap takes the reader on a journey through the research in electromagnetic wave propagation control via reconfigurable intelligent surfaces. Meta-surface modelling and design methods are reviewed along with physical realisation techniques. Several wireless applications are discussed, including beam-forming, focusing, imaging, localisation, and sensing, some rooted in novel architectures for future mobile communications networks towards 6G.

physics.class-ph↗

arXiv2021

Interplay between RIS and AI in Wireless Communications: Fundamentals, Architectures, Applications, and Open Research Problems

Future wireless communication networks are expected to fulfill the unprecedented performance requirements to support our highly digitized and globally data-driven society. Various technological challenges must be overcome to achieve our goal. Among many potential technologies, reconfigurable intelligent surface (RIS) and artificial intelligence (AI) have attracted extensive attention, thereby leading to a proliferation of studies for utilizing them in wireless communication systems. The RIS-based wireless communication frameworks and AI-enabled technologies, two of the promising technologies for the sixth-generation networks, interact and promote with each other, striving to collaboratively create a controllable, intelligent, reconfigurable, and programmable wireless propagation environment. This paper explores the road to implementing the combination of RIS and AI; specifically, integrating AI-enabled technologies into RIS-based frameworks for maximizing the practicality of RIS to facilitate the realization of smart radio propagation environments, elaborated from shallow to deep insights. We begin with the basic concept and fundamental characteristics of RIS, followed by the overview of the research status of RIS. Then, we analyze the inevitable trend of RIS to be combined with AI. In particular, we focus on recent research about RIS-based architectures embedded with AI, elucidating from the intelligent structures and systems of metamaterials to the AI-embedded RIS-assisted wireless communication systems. Finally, the challenges and potential of the topic are discussed.

cs.IT↗

arXiv2019

Compression and Acceleration of Neural Networks for Communications

Deep learning (DL) has achieved great success in signal processing and communications and has become a promising technology for future wireless communications. Existing works mainly focus on exploiting DL to improve the performance of communication systems. However, the high memory requirement and computational complexity constitute a major hurdle for the practical deployment of DL-based communications. In this article, we investigate how to compress and accelerate the neural networks (NNs) in communication systems. After introducing the deployment challenges for DL-based communication algorithms, we discuss some representative NN compression and acceleration techniques. Afterwards, two case studies for multiple-input-multiple-output (MIMO) communications, including DL-based channel state information feedback and signal detection, are presented to show the feasibility and potential of these techniques. We finally identify some challenges on NN compression and acceleration in DL-based communications and provide a guideline for subsequent research.

cs.IT↗