arXiv subjects

Liang Liu

Publications and source records attributed to Liang Liu.

At least 91 records · Page 5Linked to original sources

arXiv2025

SoccerNet 2025 Challenges Results

The SoccerNet 2025 Challenges mark the fifth annual edition of the SoccerNet open benchmarking effort, dedicated to advancing computer vision research in football video understanding. This year's challenges span four vision-based tasks: (1) Team Ball Action Spotting, focused on detecting ball-related actions in football broadcasts and assigning actions to teams; (2) Monocular Depth Estimation, targeting the recovery of scene geometry from single-camera broadcast clips through relative depth estimation for each pixel; (3) Multi-View Foul Recognition, requiring the analysis of multiple synchronized camera views to classify fouls and their severity; and (4) Game State Reconstruction, aimed at localizing and identifying all players from a broadcast video to reconstruct the game state on a 2D top-view of the field. Across all tasks, participants were provided with large-scale annotated datasets, unified evaluation protocols, and strong baselines as starting points. This report presents the results of each challenge, highlights the top-performing solutions, and provides insights into the progress made by the community. The SoccerNet Challenges continue to serve as a driving force for reproducible, open research at the intersection of computer vision, artificial intelligence, and sports. Detailed information about the tasks, challenges, and leaderboards can be found at https://www.soccer-net.org, with baselines and development kits available at https://github.com/SoccerNet.

cs.CV↗

arXiv2025

Beyond Diagonal Reconfigurable Intelligent Surface Enabled Sensing: Cramer-Rao Bound Optimization

Recently, beyond diagonal reconfigurable intelligent surface (BD-RIS) has emerged as a more flexible solution to engineer the wireless propagation channels, thanks to its non-diagonal reflecting matrix. Although the gain of the BD-RIS over the conventional RIS in communication has been revealed in many works, its gain in 6G sensing is still unknown. This motivates us to study the BD-RIS assisted sensing in this letter. Specifically, we derive the Cramer-Rao bound (CRB) for estimating the angle-of-arrival (AOA) from the target to the BD-RIS under the constraint that the BD-RIS scattering matrix is unitary. To minimize the CRB, we develop an optimization scheme based on an adaptive Riemannian steepest ascent algorithm that can satisfy the non-convex unitary constraint. Numerical results demonstrate that the proposed BD-RIS-assisted target localization method achieves superior sensing performance.

eess.SP↗

arXiv2025

Optimizing Rate-CRB Performance for Beyond Diagonal Reconfigurable Intelligent Surface Enabled ISAC

This letter considers a beyond diagonal reconfigurable intelligent surface (BD-RIS) aided integrated sensing and communication (ISAC) system, where the BD-RIS can help a multi-antenna base station (BS) serve multiple user equipments (UEs) and localize a target simultaneously. We formulate an optimization problem that designs the BS beamforming matrix and the BD-RIS scattering matrix to maximize UEs' sum rate subject to a localization Cramer-Rao bound (CRB) constraint and an additional unitary matrix constraint for the scattering matrix. Because unitary matrices form a manifold, our problem belongs to constrained manifold optimization. This letter proposes a log-barrier based Riemannian steepest ascent method to solve this problem effectively. Numerical results verify the effectiveness of our algorithm and the performance gain of the BD-RIS aided ISAC systems over the conventional RIS aided ISAC systems.

eess.SP↗

arXiv2025

CogDDN: A Cognitive Demand-Driven Navigation with Decision Optimization and Dual-Process Thinking

Mobile robots are increasingly required to navigate and interact within unknown and unstructured environments to meet human demands. Demand-driven navigation (DDN) enables robots to identify and locate objects based on implicit human intent, even when object locations are unknown. However, traditional data-driven DDN methods rely on pre-collected data for model training and decision-making, limiting their generalization capability in unseen scenarios. In this paper, we propose CogDDN, a VLM-based framework that emulates the human cognitive and learning mechanisms by integrating fast and slow thinking systems and selectively identifying key objects essential to fulfilling user demands. CogDDN identifies appropriate target objects by semantically aligning detected objects with the given instructions. Furthermore, it incorporates a dual-process decision-making module, comprising a Heuristic Process for rapid, efficient decisions and an Analytic Process that analyzes past errors, accumulates them in a knowledge base, and continuously improves performance. Chain of Thought (CoT) reasoning strengthens the decision-making process. Extensive closed-loop evaluations on the AI2Thor simulator with the ProcThor dataset show that CogDDN outperforms single-view camera-only methods by 15\%, demonstrating significant improvements in navigation accuracy and adaptability. The project page is available at https://yuehaohuang.github.io/CogDDN/.

cs.AI↗

arXiv2025

Low-latency D-MIMO Localization using Distributed Scalable Message-Passing Algorithm

Distributed MIMO and integrated sensing and communication are expected to be key technologies in future wireless systems, enabling reliable, low-latency communication and accurate localization. Dedicated localization solutions must support distributed architecture, provide scalability across different system configurations and meet strict latency requirements. We present a scalable message-passing localization method and architecture co-designed for a panel-based distributed MIMO system and network topology, in which interconnected units operate without centralized processing. This method jointly detects line-of-sight paths to distributed units from multipath measurements in dynamic scenarios, localizes the agent, and achieves very low latency. Additionally, we introduce a cycle-accurate system latency model based on implemented FPGA operations, and show important insights into processing latency and hardware utilization and system-level trade-offs. We compare our method to a multipath-based localization method and show that it can achieve similar localization performance, with wide enough distribution of array elements, while offering lower latency and computational complexity.

eess.SP↗

arXiv2025

SafeDriveRAG: Towards Safe Autonomous Driving with Knowledge Graph-based Retrieval-Augmented Generation

In this work, we study how vision-language models (VLMs) can be utilized to enhance the safety for the autonomous driving system, including perception, situational understanding, and path planning. However, existing research has largely overlooked the evaluation of these models in traffic safety-critical driving scenarios. To bridge this gap, we create the benchmark (SafeDrive228K) and propose a new baseline based on VLM with knowledge graph-based retrieval-augmented generation (SafeDriveRAG) for visual question answering (VQA). Specifically, we introduce SafeDrive228K, the first large-scale multimodal question-answering benchmark comprising 228K examples across 18 sub-tasks. This benchmark encompasses a diverse range of traffic safety queries, from traffic accidents and corner cases to common safety knowledge, enabling a thorough assessment of the comprehension and reasoning abilities of the models. Furthermore, we propose a plug-and-play multimodal knowledge graph-based retrieval-augmented generation approach that employs a novel multi-scale subgraph retrieval algorithm for efficient information retrieval. By incorporating traffic safety guidelines collected from the Internet, this framework further enhances the model's capacity to handle safety-critical situations. Finally, we conduct comprehensive evaluations on five mainstream VLMs to assess their reliability in safety-sensitive driving tasks. Experimental results demonstrate that integrating RAG significantly improves performance, achieving a +4.73% gain in Traffic Accidents tasks, +8.79% in Corner Cases tasks and +14.57% in Traffic Safety Commonsense across five mainstream VLMs, underscoring the potential of our proposed benchmark and methodology for advancing research in traffic safety. Our source code and data are available at https://github.com/Lumos0507/SafeDriveRAG.

cs.AI↗

arXiv2025

A Mousetrap: Fooling Large Reasoning Models for Jailbreak with Chain of Iterative Chaos

Large Reasoning Models (LRMs) have significantly advanced beyond traditional Large Language Models (LLMs) with their exceptional logical reasoning capabilities, yet these improvements introduce heightened safety risks. When subjected to jailbreak attacks, their ability to generate more targeted and organized content can lead to greater harm. Although some studies claim that reasoning enables safer LRMs against existing LLM attacks, they overlook the inherent flaws within the reasoning process itself. To address this gap, we propose the first jailbreak attack targeting LRMs, exploiting their unique vulnerabilities stemming from the advanced reasoning capabilities. Specifically, we introduce a Chaos Machine, a novel component to transform attack prompts with diverse one-to-one mappings. The chaos mappings iteratively generated by the machine are embedded into the reasoning chain, which strengthens the variability and complexity and also promotes a more robust attack. Based on this, we construct the Mousetrap framework, which makes attacks projected into nonlinear-like low sample spaces with mismatched generalization enhanced. Also, due to the more competing objectives, LRMs gradually maintain the inertia of unpredictable iterative reasoning and fall into our trap. Success rates of the Mousetrap attacking o1-mini, Claude-Sonnet and Gemini-Thinking are as high as 96%, 86% and 98% respectively on our toxic dataset Trotter. On benchmarks such as AdvBench, StrongREJECT, and HarmBench, attacking Claude-Sonnet, well-known for its safety, Mousetrap can astonishingly achieve success rates of 87.5%, 86.58% and 93.13% respectively. Attention: This paper contains inappropriate, offensive and harmful content.

cs.CR↗

arXiv2025

AMEX: Android Multi-annotation Expo Dataset for Mobile GUI Agents

AI agents have drawn increasing attention mostly on their ability to perceive environments, understand tasks, and autonomously achieve goals. To advance research on AI agents in mobile scenarios, we introduce the Android Multi-annotation EXpo (AMEX), a comprehensive, large-scale dataset designed for generalist mobile GUI-control agents which are capable of completing tasks by directly interacting with the graphical user interface (GUI) on mobile devices. AMEX comprises over 104K high-resolution screenshots from popular mobile applications, which are annotated at multiple levels. Unlike existing GUI-related datasets, e.g., Rico, AitW, etc., AMEX includes three levels of annotations: GUI interactive element grounding, GUI screen and element functionality descriptions, and complex natural language instructions with stepwise GUI-action chains. We develop this dataset from a more instructive and detailed perspective, complementing the general settings of existing datasets. Additionally, we finetune a baseline model SPHINX Agent and illustrate the effectiveness of AMEX.The project is available at https://yxchai.com/AMEX/.

cs.HC↗

arXiv2025

UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents

In this paper, we introduce UI-Genie, a self-improving framework addressing two key challenges in GUI agents: verification of trajectory outcome is challenging and high-quality training data are not scalable. These challenges are addressed by a reward model and a self-improving pipeline, respectively. The reward model, UI-Genie-RM, features an image-text interleaved architecture that efficiently pro- cesses historical context and unifies action-level and task-level rewards. To sup- port the training of UI-Genie-RM, we develop deliberately-designed data genera- tion strategies including rule-based verification, controlled trajectory corruption, and hard negative mining. To address the second challenge, a self-improvement pipeline progressively expands solvable complex GUI tasks by enhancing both the agent and reward models through reward-guided exploration and outcome verification in dynamic environments. For training the model, we generate UI- Genie-RM-517k and UI-Genie-Agent-16k, establishing the first reward-specific dataset for GUI agents while demonstrating high-quality synthetic trajectory gen- eration without manual annotation. Experimental results show that UI-Genie achieves state-of-the-art performance across multiple GUI agent benchmarks with three generations of data-model self-improvement. We open-source our complete framework implementation and generated datasets to facilitate further research in https://github.com/Euphoria16/UI-Genie.

cs.CL↗

arXiv2025

UI-R1: Enhancing Efficient Action Prediction of GUI Agents by Reinforcement Learning

The recent DeepSeek-R1 has showcased the emergence of reasoning capabilities in LLMs through reinforcement learning (RL) with rule-based rewards. Despite its success in language models, its application in multi-modal domains, particularly in graphic user interface (GUI) agent tasks, remains under-explored. To address this issue, we propose UI-R1, the first framework to explore how rule-based RL can enhance the reasoning capabilities of multimodal large language models (MLLMs) for GUI action prediction tasks. Specifically, UI-R1 introduces a novel rule-based action reward, enabling model optimization via policy-based algorithms such as Group Relative Policy Optimization (GRPO). For efficient training, we curate a small yet high-quality dataset of 136 challenging tasks, encompassing five common action types on mobile devices. Experimental results demonstrate that our proposed UI-R1-3B achieves significant improvements over the base model (i.e. Qwen2.5-VL-3B) on both in-domain (ID) and out-of-domain (OOD) tasks, with average accuracy gains of 22.1% on ScreenSpot, 6.0% on ScreenSpot-Pro, and 12.7% on ANDROIDCONTROL. Furthermore, UI-R1-3B delivers competitive performance compared to larger models (e.g., OS-Atlas-7B) trained via supervised fine-tuning (SFT) on 76K samples. We additionally develop an optimized version, UI-R1-E-3B, which significantly improves both grounding efficiency and accuracy. These results underscore the potential of rule-based reinforcement learning to advance GUI understanding and control, paving the way for future research in this domain. Code website: https://github.com/lll6gg/UI-R1.

cs.AI↗

arXiv2025

Adaptive Attention-Based Model for 5G Radio-based Outdoor Localization

Radio-based localization in dynamic environments, such as urban and vehicular settings, requires systems that efficiently adapt to varying signal conditions and environmental changes. Factors like multipath interference and obstructions introduce different levels of complexity that affect the accuracy of the localization. Although generalized models offer broad applicability, they often struggle to capture the nuances of specific environments, leading to suboptimal performance in real-world deployments. In contrast, specialized models can be tailored to particular conditions, enabling more precise localization by effectively handling domain-specific variations, which also results in reduced execution time and smaller model size. However, deploying multiple specialized models requires an efficient mechanism to select the most appropriate one for a given scenario. In this work, we develop an adaptive localization framework that combines shallow attention-based models with a router/switching mechanism based on a single-layer perceptron. This enables seamless transitions between specialized localization models optimized for different conditions, balancing accuracy and computational complexity. We design three low-complex models tailored for distinct scenarios, and a router that dynamically selects the most suitable model based on real-time input characteristics. The proposed framework is validated using real-world vehicle localization data collected from a massive MIMO base station and compared to more general models.

eess.SP↗

arXiv2025

ESTA: An Efficient Spatial-Temporal Range Aggregation Query Processing Algorithm for UAV Networks

Unmanned Aerial Vehicle (UAV) networks are increasingly deployed in military and civilian applications, serving as critical platforms for data collection. Users frequently require aggregated statistical information derived from historical sensory data within specific spatial and temporal boundaries. To address this, users submit aggregation query requests with spatial-temporal constraints to target UAVs that store the relevant data. These UAVs process and return the query results, which can be aggregated within the network during transmission to conserve energy and bandwidth-resources that are inherently limited in UAV networks. However,the dynamic topology caused by UAV mobility, coupled with these resource constraints, makes efficient in-network aggregation challenging without compromising user query delay. To the best of our knowledge, existing research has yet to adequately explore spatial-temporal range aggregation queries in the context of UAV networks. In this paper, we propose ESTA, an Efficient Spatial-Temporal range Aggregation query processing algorithm tailored for UAV networks. ESTA leverages pre-planned UAV trajectories to construct a topology change graph that models the network's evolving connectivity. It then employs an efficient shortest path algorithm to determine the minimum query response delay. Subsequently, while adhering to user-specified delay constraints, ESTA transforms the in-network aggregation process into a series of set cover problems, which are solved recursively to build a Spatial-Temporal Aggregation Tree (STAT). This tree enables the identification of an energy-efficient routing path for aggregating and delivering query results. Extensive simulations demonstrate that ESTA reduces energy consumption by more than 50% compared to a baseline algorithm, all while satisfying the required query delay.

cs.DC↗

arXiv2025

Search for New Particles with Flying Quantum Sensors in Space

Recent advancements in space science and technologies offer exciting prospects for investigating novel research that is unattainable within terrestrial laboratories. Here we propose the implementation of space-based quantum sensing to explore ultralight new particles beyond the standard model. The central idea involves probing long-range interactions between spin ensembles of space quantum sensors and the particles residing within Earth, mediated by ultralight particles. We show that such interactions can be substantially enhanced in space platforms and thus increase the search sensitivity. In contrast to their terrestrial counterparts, space-based quantum searches exhibit remarkable velocity enhancements, approaching the first cosmic speed, and thus enables the exploration of unexplored parameter space concerning ultralight new particles. Furthermore, the substantial abundance of electrons and nucleons within Earth plays a crucial role in extending the scope of our mission. Our projected search sensitivity can surpass the sensitivities of both terrestrial experiments and proposals by up to approximately 7 orders of magnitude. We also briefly discuss other space mission, including ``space-ground integrated" network of quantum sensors for dark matter searches.

quant-ph↗

arXiv2025

QoS-based Beamforming and Compression Design for Cooperative Cellular Networks via Lagrangian Duality

This paper considers the quality-of-service (QoS)-based joint beamforming and compression design problem in the downlink cooperative cellular network, where multiple relay-like base stations (BSs), connected to the central processor via rate-limited fronthaul links, cooperatively transmit messages to the users. The problem of interest is formulated as the minimization of the total transmit power of the BSs, subject to all users' signal-to-interference-plus-noise ratio (SINR) constraints and all BSs' fronthaul rate constraints. In this paper, we first show that there is no duality gap between the considered joint optimization problem and its Lagrangian dual by showing the tightness of its semidefinite relaxation (SDR). Then, we propose an efficient algorithm based on the above duality result for solving the considered problem. The proposed algorithm judiciously exploits the special structure of an enhanced Karush-Kuhn-Tucker (KKT) conditions of the considered problem and finds the solution that satisfies the enhanced KKT conditions via two fixed point iterations. Two key features of the proposed algorithm are: (1) it is able to detect whether the considered problem is feasible or not and find its globally optimal solution when it is feasible; (2) it is highly efficient because both of the fixed point iterations in the proposed algorithm are linearly convergent and evaluating the functions in the fixed point iterations are computationally cheap. Numerical results show the global optimality and efficiency of the proposed algorithm.

cs.IT↗

arXiv2025

T2SG: Traffic Topology Scene Graph for Topology Reasoning in Autonomous Driving

Understanding the traffic scenes and then generating high-definition (HD) maps present significant challenges in autonomous driving. In this paper, we defined a novel Traffic Topology Scene Graph, a unified scene graph explicitly modeling the lane, controlled and guided by different road signals (e.g., right turn), and topology relationships among them, which is always ignored by previous high-definition (HD) mapping methods. For the generation of T2SG, we propose TopoFormer, a novel one-stage Topology Scene Graph TransFormer with two newly designed layers. Specifically, TopoFormer incorporates a Lane Aggregation Layer (LAL) that leverages the geometric distance among the centerline of lanes to guide the aggregation of global information. Furthermore, we proposed a Counterfactual Intervention Layer (CIL) to model the reasonable road structure ( e.g., intersection, straight) among lanes under counterfactual intervention. Then the generated T2SG can provide a more accurate and explainable description of the topological structure in traffic scenes. Experimental results demonstrate that TopoFormer outperforms existing methods on the T2SG generation task, and the generated T2SG significantly enhances traffic topology reasoning in downstream tasks, achieving a state-of-the-art performance of 46.3 OLS on the OpenLane-V2 benchmark. We will release our source code and model.

cs.CV↗

arXiv2025

LearnAct: Few-Shot Mobile GUI Agent with a Unified Demonstration Benchmark

Mobile GUI agents show promise in automating tasks but face generalization challenges in diverse real-world scenarios. Traditional approaches using pre-training or fine-tuning with massive datasets struggle with the diversity of mobile applications and user-specific tasks. We propose enhancing mobile GUI agent capabilities through human demonstrations, focusing on improving performance in unseen scenarios rather than pursuing universal generalization through larger datasets. To realize this paradigm, we introduce LearnGUI, the first comprehensive dataset specifically designed for studying demonstration-based learning in mobile GUI agents, comprising 2,252 offline tasks and 101 online tasks with high-quality human demonstrations. We further develop LearnAct, a sophisticated multi-agent framework that automatically extracts knowledge from demonstrations to enhance task completion. This framework integrates three specialized agents: DemoParser for knowledge extraction, KnowSeeker for relevant knowledge retrieval, and ActExecutor for demonstration-enhanced task execution. Our experimental results show significant performance gains in both offline and online evaluations. In offline assessments, a single demonstration improves model performance, increasing Gemini-1.5-Pro's accuracy from 19.3% to 51.7%. In online evaluations, our framework enhances UI-TARS-7B-SFT's task success rate from 18.1% to 32.8%. LearnAct framework and LearnGUI benchmark establish demonstration-based learning as a promising direction for more adaptable, personalized, and deployable mobile GUI agents.

cs.HC↗

arXiv2025

Over-the-Air Diagnosis of Defective Elements in Intelligent Reflecting Surface

Due to circuit failures, defective elements that cannot adaptively adjust the phase shifts of their impinging signals in a desired manner may exist on an intelligent reflecting surface (IRS). Traditional way to locate these defective IRS elements requires a thorough diagnosis of all the circuits belonging to a huge number of IRS elements, which is practically challenging. In this paper, we will devise novel approaches under which a transmitter sends known pilot signals and a receiver localizes all the defective IRS elements just based on its over-the-air measurements reflected from the IRS. Specifically, given any set of IRS elements, we propose an efficient method to process the received signals to determine whether this cluster contains defective elements or not with a very high accuracy probability. Based on this method, we show that the over-the-air diagnosis problem belongs to the 20 questions problem, where we can adaptively change the query set at the IRS so as to localize all the defective elements as quickly as possible. Along this line, we first propose a sorted posterior matching (sortPM) based method according to the noisy 20 questions technique, which enables accurate diagnosis even if the answers about the existence of defective elements in some sets of interest are wrong at certain question and answer (Q&A) rounds due to the noisy received signals. Next, to reduce the complexity, we propose a bisection based method according to the noiseless 20 questions technique, which totally trusts the answer at each Q&A round and keeps removing half of the remaining region based on such answers. Via numerical results, we show that our proposed methods can exploit the over-the-air measurements to localize all the defective IRS elements quickly and accurately.

eess.SP↗

arXiv2025

Low-Overhead Channel Estimation Framework for Beyond Diagonal Reconfigurable Intelligent Surface Assisted Multi-User MIMO Communication

Beyond diagonal reconfigurable intelligent surface (BD-RIS) refers to a family of RIS architectures characterized by scattering matrices not limited to being diagonal and enables higher wave manipulation flexibility and large performance gains over conventional (diagonal) RIS. To achieve those promising gains, accurate channel state information (CSI) needs to be acquired in BD-RIS assisted communication systems. However, the number of coefficients in the cascaded channels to be estimated in BD-RIS assisted systems is significantly larger than that in conventional RIS assisted systems, because the channels associated with the off-diagonal elements of the scattering matrix have to be estimated as well. Surprisingly, for the first time in the literature, this paper rigorously shows that the uplink channel estimation overhead in BD-RIS assisted systems is actually of the same order as that in the conventional RIS assisted systems. This amazing result stems from a key observation: for each user antenna, its cascaded channel matrix associated with one reference BD-RIS element is a scaled version of that associated with any other BD-RIS element due to the common RIS-base station (BS) channel. In other words, the number of independent unknown variables is far less than it would seem at first glance. Building upon this property, this paper manages to characterize the minimum overhead to perfectly estimate all the channels in the ideal case without noise at the BS, and propose a twophase estimation framework for the practical case with noise at the BS. Numerical results demonstrate outstanding channel estimation overhead reduction over existing schemes in BD-RIS assisted systems.

eess.SP↗