arXiv subjects

William Wang

Publications and source records attributed to William Wang.

At least 19 recordsLinked to original sources

arXiv2026

Learning Realistic Athletic Sprinting Without Demonstrations

We present a muscle-driven simulation system for generating biomechanically accurate motion for high-speed athletic locomotion tasks that does not require motion demonstrations. Our approach integrates state-of-the-art biomechanical athlete models into a new, high-performance GPU simulator capable of running at 1000x real-time. High-throughput simulation enables large-batch reinforcement learning to train control policies that operate directly in the model's high-dimensional muscle excitation space, and are guided only by task-specific episode termination conditions and a reward that encourages maximizing speed while reducing forces needed to respect joint limits. These policies train within a few hours on a single GPU and generate "near visually realistic" motions for complete athletic activities such as a full 100-meter sprint or performing popular athletic locomotion drills like side-shuffling, backpedaling, and carioca. The generated sprinting motions also exhibit strong agreement with experimental data captured from sprinters.

cs.GR↗

arXiv2026

Magnetic Field Reorganization of Electronic States in Moir\'e Bilayer Graphene

Magnetic fields are widely used to diagnose quantum phases in two-dimensional systems through quantum oscillations or by tuning spin and valley polarizations, but magnetic fields can also reshape the underlying electronic structure. Here, using a bilayer graphene/hBN moir\'e system, we reveal a rich magnetic field induced evolution of the semiclassical orbit network, encompassing Lifshitz transitions, magnetic breakdown, and scattering between coexisting electron and hole pockets. At magnetic fields of 1 T to 2 T, quantum oscillation frequencies and the Hall density change markedly over a broad carrier density range, signaling magnetic breakdown and magnetic Lifshitz transitions. This evolution is valley contrasting: Berry curvature hot spots near the breakdown junctions enhance magnetic breakdown in the K valley while suppressing it in the K$^\prime$ valley, whereas valley-antisymmetric orbital magnetic moments split the corresponding Lifshitz transitions. The resulting valley-selective trajectories manifest at higher fields as valley-symmetry-breaking Hofstadter gaps. At elevated temperatures and low magnetic fields, scattering between coexisting electron and hole pockets produces nearly density-independent resistance oscillations whose frequency tracks the sum of their Fermi surface areas, persisting after conventional Onsager oscillations are thermally washed out. Our results provide a unified picture of how modest magnetic fields reorganize moir\'e electronic states as the system evolves from semiclassical transport toward the Hofstadter regime.

cond-mat.mes-hall↗

arXiv2026

Flock: Fast Proving for Batch Boolean Computations

For many applications of SNARKs, a key bottleneck is proving large batches of standard cryptographic hash evaluations, such as SHA-256, Keccak, or BLAKE3. We introduce Flock, a hash-based SNARK for extremely fast proving of such batched Boolean computations. Flock proves batches of the same R1CS circuit (plus input/output relations between them), can prove hash-chains and Merkle path openings, and in principle can be extended to full-fledged hash-based signature verification. At its core, Flock combines new optimizations for the lincheck and zerocheck protocols with an aggressively optimized proof-of-concept implementation co-designed by coding agents. On a single core of an M4 Max processor, Flock proves 82k evaluations of the BLAKE3 compression function, 42k SHA-256 compressions, and 30k Keccak permutations per second --- less than a $250\times$ overhead over native execution. On ten cores, throughput exceeds 660k BLAKE3 compressions per second; in proving SHA-256, Flock is more than $9\times$ faster than Binius64, the prior state of the art, and more than $500\times$ faster than the fastest elliptic curve-based SNARK we measured against.

cs.CR↗

arXiv2026

The MICADO first light imager for the ELT: the SCAO module AIT in France, from the beta flat configuration towards the flat and final configurations

MICADO is the designated ELT first light instrument, a near-infrared imager working at the diffraction limit of the telescope thanks to a SCAO and an MCAO mode. Since summer 2024, MICADO is in its AIT phase and all MICADO partners are currently integrating their subsystems. The MICADO SCAO module will follow three successive AIT phases in France: the beta flat configuration, the flat configuration and the final configuration. In this contribution I will first present the progress made in the AIT of the beta flat configuration and its closed loop results. I will then present the performed and on-going tasks for the flat and final configurations (manufacturing, delivery and integration of several SCAO subsystems and pieces of software) and the results obtained so far in the flat configuration. I will then conclude with the expected following steps of the SCAO module AIT in Garching.

astro-ph.IM↗

arXiv2026

"Are you an AI?" Analyzing Client Suspicion of AI Use in Crisis Counseling

As artificial intelligence (AI) tools get increasingly deployed for mental healthcare, public trust in these systems remains uncertain. It is unclear how clients perceive AI involvement in counseling interactions, particularly in moments of crisis that require empathy and connection. To address this gap, we analyzed 75,777 crisis counseling conversations from a human-staffed WhatsApp helpline in India to characterize how often clients suspected they were speaking to AI, what triggered those doubts, and how counselors responded. Though no conversations actually involved AI assistance, the proportion of conversations where clients suspected AI use increased from 0.8% in June 2024 to 2.6% in March 2025. Within suspicious conversations, 21.5% of clients stated an explicit preference for humans. Client suspicion primarily arose in the first half of messages (68.3%), and when counselors offered reassurance (e.g. 'I assure you; this is not ai!'), clients continued to press or ended the conversation 17.6% of the time. As AI tools get increasingly integrated into counselor workflows, understanding these dynamics is essential for designing AI systems that preserve the therapeutic relationship between counselors and clients.

cs.HC↗

arXiv2026

Planning to Explore: Curiosity-Driven Planning for LLM Test Generation

The use of LLMs for code generation has naturally extended to code testing and evaluation. As codebases grow in size and complexity, so does the need for automated test generation. Current approaches for LLM-based test generation rely on strategies that maximize immediate coverage gain, a greedy approach that plateaus on code where reaching deep branches requires setup steps that individually yield zero new coverage. Drawing on principles of Bayesian exploration, we treat the program's branch structure as an unknown environment, and an evolving coverage map as a proxy probabilistic posterior representing what the LLM has discovered so far. Our method, CovQValue, feeds the coverage map back to the LLM, generates diverse candidate plans in parallel, and selects the most informative plan by LLM-estimated Q-values, seeking actions that balance immediate branch discovery with future reachability. Our method outperforms greedy selection on TestGenEval Lite, achieving 51-77% higher branch coverage across three popular LLMs and winning on 77-84% of targets. In addition, we build a benchmark for iterative test generation, RepoExploreBench, where they achieve 40-74%. These results show the potential of curiosity-driven planning methods for LLM-based exploration, enabling more effective discovery of program behavior through sequential interaction

cs.SE↗

arXiv2025

Univariate-Guided Sparse Regression for Biobank-Scale High-Dimensional Omics Data

We present a scalable framework for computing polygenic risk scores (PRS) in high-dimensional genomic settings using the recently introduced Univariate-Guided Sparse Regression (uniLasso). UniLasso is a two-stage penalized regression procedure that leverages univariate coefficients and magnitudes to stabilize feature selection and enhance interpretability. Building on its theoretical and empirical advantages, we adapt uniLasso for application to the UK Biobank, a population-based repository comprising over one million genetic variants measured on hundreds of thousands of individuals from the United Kingdom. We further extend the framework to incorporate external summary statistics to increase predictive accuracy. Our results demonstrate that uniLasso attains predictive performance comparable to standard Lasso while selecting substantially fewer variants, yielding sparser and more interpretable models. Moreover, it exhibits superior performance in estimating PRS relative to its competitors, such as PRS-CS. Integrating external scores further improves prediction while maintaining sparsity.

stat.ME↗

arXiv2025

SpikeATac: A Multimodal Tactile Finger with Taxelized Dynamic Sensing for Dexterous Manipulation

In this work, we introduce SpikeATac, a multimodal tactile finger combining a taxelized and highly sensitive dynamic response (PVDF) with a static transduction method (capacitive) for multimodal touch sensing. Named for its `spiky' response, SpikeATac's 16-taxel PVDF film sampled at 4 kHz provides fast, sensitive dynamic signals to the very onset and breaking of contact. We characterize the sensitivity of the different modalities, and show that SpikeATac provides the ability to stop quickly and delicately when grasping fragile, deformable objects. Beyond parallel grasping, we show that SpikeATac can be used in a learning-based framework to achieve new capabilities on a dexterous multifingered robot hand. We use reinforcement learning from human feedback to fine-tune the behavior of a policy to modulate force. Our hardware platform and learning pipeline together enable a difficult dexterous and contact-rich task that has not previously been achieved: in-hand manipulation of fragile objects. Videos are available at https://roamlab.github.io/spikeatac/ .

cs.RO↗

arXiv2025

Throttling for metric dimension and its variants

Metric dimension is a graph parameter that has been applied to robot navigation and finding low-dimensional vector embeddings. Throttling entails minimizing the sum of two available resources when solving certain graph problems. In this paper, we introduce throttling for metric dimension, edge metric dimension, and mixed metric dimension. In the context of vector embeddings, metric dimension throttling finds a low-dimensional, low-magnitude embedding with integer coordinates. We show that computing the throttling number is NP-hard for all three variants. We give formulas for the throttling numbers of special families of graphs, and characterize graphs with extremal throttling numbers. We also prove that the minimum possible throttling number of a graph of order $n$ is $\Theta\left(\frac{\log{n}}{\log{\log{n}}}\right)$, while the minimum possible throttling number of a tree of order $n$ is $\Theta(n^{1/3})$ or $\Theta(n^{1/2})$ depending on the variant of metric dimension.

math.CO↗

arXiv2025

Agents of Change: Self-Evolving LLM Agents for Strategic Planning

We address the long-horizon gap in large language model (LLM) agents by enabling them to sustain coherent strategies in adversarial, stochastic environments. Settlers of Catan provides a challenging benchmark: success depends on balancing short- and long-term goals amid randomness, trading, expansion, and blocking. Prompt-centric LLM agents (e.g., ReAct, Reflexion) must re-interpret large, evolving game states each turn, quickly saturating context windows and losing strategic consistency. We propose HexMachina, a continual learning multi-agent system that separates environment discovery (inducing an adapter layer without documentation) from strategy improvement (evolving a compiled player through code refinement and simulation). This design preserves executable artifacts, allowing the LLM to focus on high-level strategy rather than per-turn reasoning. In controlled Catanatron experiments, HexMachina learns from scratch and evolves players that outperform the strongest human-crafted baseline (AlphaBeta), achieving a 54% win rate and surpassing prompt-driven and no-discovery baselines. Ablations confirm that isolating pure strategy learning improves performance. Overall, artifact-centric continual learning transforms LLMs from brittle stepwise deciders into stable strategy designers, advancing long-horizon autonomy.

cs.AI↗

arXiv2025

How Is LLM Reasoning Distracted by Irrelevant Context? An Analysis Using a Controlled Benchmark

We introduce Grade School Math with Distracting Context (GSM-DC), a synthetic benchmark to evaluate Large Language Models' (LLMs) reasoning robustness against systematically controlled irrelevant context (IC). GSM-DC constructs symbolic reasoning graphs with precise distractor injections, enabling rigorous, reproducible evaluation. Our experiments demonstrate that LLMs are significantly sensitive to IC, affecting both reasoning path selection and arithmetic accuracy. Additionally, training models with strong distractors improves performance in both in-distribution and out-of-distribution scenarios. We further propose a stepwise tree search guided by a process reward model, which notably enhances robustness in out-of-distribution conditions.

cs.CL↗

arXiv2025

Self-Resource Allocation in Multi-Agent LLM Systems

With the development of LLMs as agents, there is a growing interest in connecting multiple agents into multi-agent systems to solve tasks concurrently, focusing on their role in task assignment and coordination. This paper explores how LLMs can effectively allocate computational tasks among multiple agents, considering factors such as cost, efficiency, and performance. In this work, we address key questions, including the effectiveness of LLMs as orchestrators and planners, comparing their effectiveness in task assignment and coordination. Our experiments demonstrate that LLMs can achieve high validity and accuracy in resource allocation tasks. We find that the planner method outperforms the orchestrator method in handling concurrent actions, resulting in improved efficiency and better utilization of agents. Additionally, we show that providing explicit information about worker capabilities enhances the allocation strategies of planners, particularly when dealing with suboptimal workers.

cs.MA↗

arXiv2025

VITED: Video Temporal Evidence Distillation

We investigate complex video question answering via chain-of-evidence reasoning -- identifying sequences of temporal spans from multiple relevant parts of the video, together with visual evidence within them. Existing models struggle with multi-step reasoning as they uniformly sample a fixed number of frames, which can miss critical evidence distributed nonuniformly throughout the video. Moreover, they lack the ability to temporally localize such evidence in the broader context of the full video, which is required for answering complex questions. We propose a framework to enhance existing VideoQA datasets with evidence reasoning chains, automatically constructed by searching for optimal intervals of interest in the video with supporting evidence, that maximizes the likelihood of answering a given question. We train our model (VITED) to generate these evidence chains directly, enabling it to both localize evidence windows as well as perform multi-step reasoning across them in long-form video content. We show the value of our evidence-distilled models on a suite of long video QA benchmarks where we outperform state-of-the-art approaches that lack evidence reasoning capabilities.

cs.CV↗

arXiv2024

Game-theoretic LLM: Agent Workflow for Negotiation Games

This paper investigates the rationality of large language models (LLMs) in strategic decision-making contexts, specifically within the framework of game theory. We evaluate several state-of-the-art LLMs across a spectrum of complete-information and incomplete-information games. Our findings reveal that LLMs frequently deviate from rational strategies, particularly as the complexity of the game increases with larger payoff matrices or deeper sequential trees. To address these limitations, we design multiple game-theoretic workflows that guide the reasoning and decision-making processes of LLMs. These workflows aim to enhance the models' ability to compute Nash Equilibria and make rational choices, even under conditions of uncertainty and incomplete information. Experimental results demonstrate that the adoption of these workflows significantly improves the rationality and robustness of LLMs in game-theoretic tasks. Specifically, with the workflow, LLMs exhibit marked improvements in identifying optimal strategies, achieving near-optimal allocations in negotiation scenarios, and reducing susceptibility to exploitation during negotiations. Furthermore, we explore the meta-strategic considerations of whether it is rational for agents to adopt such workflows, recognizing that the decision to use or forgo the workflow constitutes a game-theoretic issue in itself. Our research contributes to a deeper understanding of LLMs' decision-making capabilities in strategic contexts and provides insights into enhancing their rationality through structured workflows. The findings have implications for the development of more robust and strategically sound AI agents capable of navigating complex interactive environments. Code and data supporting this study are available at \url{https://github.com/Wenyueh/game_theory}.

cs.AI↗

arXiv2024

SWE-Search: Enhancing Software Agents with Monte Carlo Tree Search and Iterative Refinement

Software engineers operating in complex and dynamic environments must continuously adapt to evolving requirements, learn iteratively from experience, and reconsider their approaches based on new insights. However, current large language model (LLM)-based software agents often follow linear, sequential processes that prevent backtracking and exploration of alternative solutions, limiting their ability to rethink their strategies when initial approaches prove ineffective. To address these challenges, we propose SWE-Search, a multi-agent framework that integrates Monte Carlo Tree Search (MCTS) with a self-improvement mechanism to enhance software agents' performance on repository-level software tasks. SWE-Search extends traditional MCTS by incorporating a hybrid value function that leverages LLMs for both numerical value estimation and qualitative evaluation. This enables self-feedback loops where agents iteratively refine their strategies based on both quantitative numerical evaluations and qualitative natural language assessments of pursued trajectories. The framework includes a SWE-Agent for adaptive exploration, a Value Agent for iterative feedback, and a Discriminator Agent that facilitates multi-agent debate for collaborative decision-making. Applied to the SWE-bench benchmark, our approach demonstrates a 23% relative improvement in performance across five models compared to standard open-source agents without MCTS. Our analysis reveals how performance scales with increased inference-time compute through deeper search, providing a pathway to improve software agents without requiring larger models or additional training data. This highlights the potential of self-evaluation driven search techniques in complex software engineering environments.

cs.AI↗

arXiv2024

A Mathematical Framework and a Suite of Learning Techniques for Neural-Symbolic Systems

The field of Neural-Symbolic (NeSy) systems is growing rapidly. Proposed approaches show great promise in achieving symbiotic unions of neural and symbolic methods. However, a unifying framework is needed to organize common NeSy modeling patterns and develop general learning approaches. In this paper, we introduce Neural-Symbolic Energy-Based Models (NeSy-EBMs), a unifying mathematical framework for discriminative and generative NeSy modeling. Importantly, NeSy-EBMs allow the derivation of general expressions for gradients of prominent learning losses, and we introduce a suite of four learning approaches that leverage methods from multiple domains, including bilevel and stochastic policy optimization. Finally, we ground the NeSy-EBM framework with Neural Probabilistic Soft Logic (NeuPSL), an open-source NeSy-EBM library designed for scalability and expressivity, facilitating the real-world application of NeSy systems. Through extensive empirical analysis across multiple datasets, we demonstrate the practical advantages of NeSy-EBMs in various tasks, including image classification, graph node labeling, autonomous vehicle situation awareness, and question answering.

cs.LG↗

arXiv2024

MultiAgent Collaboration Attack: Investigating Adversarial Attacks in Large Language Model Collaborations via Debate

Large Language Models (LLMs) have shown exceptional results on current benchmarks when working individually. The advancement in their capabilities, along with a reduction in parameter size and inference times, has facilitated the use of these models as agents, enabling interactions among multiple models to execute complex tasks. Such collaborations offer several advantages, including the use of specialized models (e.g. coding), improved confidence through multiple computations, and enhanced divergent thinking, leading to more diverse outputs. Thus, the collaborative use of language models is expected to grow significantly in the coming years. In this work, we evaluate the behavior of a network of models collaborating through debate under the influence of an adversary. We introduce pertinent metrics to assess the adversary's effectiveness, focusing on system accuracy and model agreement. Our findings highlight the importance of a model's persuasive ability in influencing others. Additionally, we explore inference-time methods to generate more compelling arguments and evaluate the potential of prompt-based mitigation as a defensive strategy.

cs.CL↗

arXiv2024

Don't Fine-Tune, Decode: Syntax Error-Free Tool Use via Constrained Decoding

Instruction-tuned large language models (LLMs) excel at many tasks but often fail to use external tools due to complicated and unfamiliar syntax constraints. While extensive fine-tuning and prompting can mitigate the issue, these approaches are expensive and hard to generalize. Furthermore, because syntax constraints are only learned implicitly during fine-tuning, models still make frequent syntax errors. Motivated by the fact that these constraints can be better satisfied explicitly with constrained decoding, we propose TOOLDEC, a decoding algorithm using finite state machines to force LLMs to follow tool syntax. Our experiments show that TOOLDEC eliminates all syntax errors, achieving significantly better performance on various base models and benchmarks. More surprisingly, when applied to generalist out-of-the-box LLMs such as Mistral-Instruct, TOOLDEC improves its accuracy in tool use from the initial 0% to an impressive 52%, matching the performance of specialized fine-tuned models such as ToolLLM.

cs.CL↗