Searcharxiv⌕ Search

arXiv subjects

Guanyu Li

Publications and source records attributed to Guanyu Li.

At least 19 recordsLinked to original sources
arXiv2026

Calibrated Pressure-Observable Born and Hessian Actions for Quantum-Assisted Waveform Inversion

We construct a pressure-consistent operator-and-readout interface for Born, adjoint, and Gauss--Newton actions in constant-density acoustic full-waveform inversion (FWI) using Schr\"odingerised propagation. The energy variables $\pi=c^{-1}\partial_t u$ and $q=\nabla u$ yield an auxiliary-space Hamiltonian, while physical pressure $p=c\pi$ depends explicitly on wavespeed. Its derivative $D(c\pi)[c_0](\delta c)=c_0\delta\pi+\delta c\,\pi_0$ combines propagated wavefield sensitivity with a direct receiver-calibration term. Duhamel and receiver-row differentiation retain both contributions in the Born map, its adjoint, and the Gauss--Newton normal action. We prove a conditional consistency estimate with a periodic second-order finite-difference specialization and give a resource model for state preparation, normalization, quadrature, and selected-output measurement. A compiled nine-qubit instance realizes structured preparation, product-formula propagation, a derivative-LCU block, and calibrated pressure-overlap measurements. Bernoulli samples from ideal-circuit probabilities drive a four-parameter hybrid inversion. A two-qubit VQLS circuit represents the normalized update direction, while normal-system assembly, line search, and model refresh remain classical. Finite differences, tangent and reverse-adjoint recurrences, autodiff JVP/VJP evaluations, and explicit Jacobians verify the discrete Born, adjoint, and normal actions. Smooth periodic refinement confirms second-order convergence, whereas omitting receiver calibration leaves an order-one Born error and substantially changes the regularized Gauss--Newton direction. All ten predeclared finite-shot runs reduce the initial model error. These results specify the physical-pressure derivative and selected-output measurements needed to connect Schr\"odingerised propagation to a local FWI update.

quant-ph↗
arXiv2026

Derived representation schemes with arbitrary coefficients and associative smoothness

We study associative smoothness through representation homology with coefficients in arbitrary finite-dimensional algebras. We prove that the higher representation homology of a finitely generated formally smooth algebra with arbitrary finite-dimensional coefficients vanishes. We further show that allowing arbitrary coefficients yields a strictly stronger smoothness test: suitable non-matrix coefficients detect nonsmoothness in situations where matrix coefficients do not. These results raise the question of whether representation homology with arbitrary coefficients furnishes a homotopical characterization of associative smoothness in the spirit of the derived Kontsevich-Rosenberg principle. For finite-dimensional algebras, we prove that representation homology with arbitrary coefficients completely characterizes formal smoothness, providing supporting evidence for this philosophy.

math.AG↗
arXiv2026

Orlov's functors in Macaulay2

Given a commutative and graded Gorenstein ring $R$ with associated projective variety $X$, a theorem of Orlov gives fully faithful embeddings from the graded singularity category of $R$ to the derived category of $X$, or vice versa, depending on the degree of the canonical bundle of $X$. We describe algorithms for computing these embeddings that can be implemented in Macaulay2.

math.AC↗
arXiv2026

Kimi K2.5: Visual Agentic Intelligence

We introduce Kimi K2.5, an open-source multimodal agentic model designed to advance general agentic intelligence. K2.5 emphasizes the joint optimization of text and vision so that two modalities enhance each other. This includes a series of techniques such as joint text-vision pre-training, zero-vision SFT, and joint text-vision reinforcement learning. Building on this multimodal foundation, K2.5 introduces Agent Swarm, a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. Extensive evaluations show that Kimi K2.5 achieves state-of-the-art results across various domains including coding, vision, reasoning, and agentic tasks. Agent Swarm also reduces latency by up to $4.5\times$ over single-agent baselines. We release the post-trained Kimi K2.5 model checkpoint to facilitate future research and real-world applications of agentic intelligence.

cs.CL↗
arXiv2025

AgentPRM: Process Reward Models for LLM Agents via Step-Wise Promise and Progress

Despite rapid development, large language models (LLMs) still encounter challenges in multi-turn decision-making tasks (i.e., agent tasks) like web shopping and browser navigation, which require making a sequence of intelligent decisions based on environmental feedback. Previous work for LLM agents typically relies on elaborate prompt engineering or fine-tuning with expert trajectories to improve performance. In this work, we take a different perspective: we explore constructing process reward models (PRMs) to evaluate each decision and guide the agent's decision-making process. Unlike LLM reasoning, where each step is scored based on correctness, actions in agent tasks do not have a clear-cut correctness. Instead, they should be evaluated based on their proximity to the goal and the progress they have made. Building on this insight, we propose a re-defined PRM for agent tasks, named AgentPRM, to capture both the interdependence between sequential decisions and their contribution to the final goal. This enables better progress tracking and exploration-exploitation balance. To scalably obtain labeled data for training AgentPRM, we employ a Temporal Difference-based (TD-based) estimation method combined with Generalized Advantage Estimation (GAE), which proves more sample-efficient than prior methods. Extensive experiments across different agentic tasks show that AgentPRM is over $8\times$ more compute-efficient than baselines, and it demonstrates robust improvement when scaling up test-time compute. Moreover, we perform detailed analyses to show how our method works and offer more insights, e.g., applying AgentPRM to the reinforcement learning of LLM agents.

cs.CL↗
arXiv2025

Computing global Ext for complexes

We give a computational algorithm for computing Ext groups between bounded complexes of coherent sheaves on a projective variety, and we describe an implementation of this algorithm in Macaulay2. In particular, our results yield methods for computing derived global sections of bounded complexes of coherent sheaves and mutations of exceptional collections.

math.AG↗
arXiv2025

AgentGym-RL: Training LLM Agents for Long-Horizon Decision Making through Multi-Turn Reinforcement Learning

Developing autonomous LLM agents capable of making a series of intelligent decisions to solve complex, real-world tasks is a fast-evolving frontier. Like human cognitive development, agents are expected to acquire knowledge and skills through exploration and interaction with the environment. Despite advances, the community still lacks a unified, interactive reinforcement learning (RL) framework that can effectively train such agents from scratch -- without relying on supervised fine-tuning (SFT) -- across diverse and realistic environments. To bridge this gap, we introduce AgentGym-RL, a new framework to train LLM agents for multi-turn interactive decision-making through RL. The framework features a modular and decoupled architecture, ensuring high flexibility and extensibility. It encompasses a wide variety of real-world scenarios, and supports mainstream RL algorithms. Furthermore, we propose ScalingInter-RL, a training approach designed for exploration-exploitation balance and stable RL optimization. In early stages, it emphasizes exploitation by restricting the number of interactions, and gradually shifts towards exploration with larger horizons to encourage diverse problem-solving strategies. In this way, the agent develops more diverse behaviors and is less prone to collapse under long horizons. We perform extensive experiments to validate the stability and effectiveness of both the AgentGym-RL framework and the ScalingInter-RL approach. Our agents match or surpass commercial models on 27 tasks across diverse environments. We offer key insights and will open-source the complete AgentGym-RL framework -- including code and datasets -- to empower the research community in developing the next generation of intelligent agents.

cs.LG↗
arXiv2025

BMMR: A Large-Scale Bilingual Multimodal Multi-Discipline Reasoning Dataset

In this paper, we introduce BMMR, a large-scale bilingual, multimodal, multi-disciplinary reasoning dataset for the community to develop and evaluate large multimodal models (LMMs). BMMR comprises 110k college-level questions spanning 300 UNESCO-defined subjects, spanning diverse formats-multiple-choice, fill-in-the-blank, and open-ended QA-and sourced from both print and digital media such as books, exams, and quizzes. All data are curated and filtered via a human-in-the-loop and scalable framework, and each instance is paired with a high-quality reasoning path. The dataset is organized into two parts: BMMR-Eval that comprises 20,458 high-quality instances to comprehensively assess LMMs' knowledge and reasoning across multiple disciplines in both Chinese and English; and BMMR-Train that contains 88,991 instances to support further research and development, extending the current focus on mathematical reasoning to diverse disciplines and domains. In addition, we propose the process-based multi-discipline verifier (i.e., BMMR-Verifier) for accurate and fine-grained evaluation of reasoning paths. Extensive experiments on 24 models reveal that (i) even SOTA models (e.g., o3 and Gemini-2.5-Pro) leave substantial headroom on BMMR-Eval; (ii) reasoning models exhibit discipline bias and outperform LMMs only on specific subjects; (iii) open-source models still trail their proprietary counterparts; and (iv) fine-tuning on BMMR-Train narrows this gap. Additionally, we conduct reasoning-chain analyses using BMMR-Verifier and other in-depth studies, uncovering the challenges LMMs currently face in multidisciplinary reasoning. We will release the data, and we hope our work can offer insights and contributions to the community.

cs.CL↗
arXiv2024

Enhancing LLM Reasoning via Critique Models with Test-Time and Training-Time Supervision

Training large language models (LLMs) to spend more time thinking and reflection before responding is crucial for effectively solving complex reasoning tasks in fields such as science, coding, and mathematics. However, the effectiveness of mechanisms like self-reflection and self-correction depends on the model's capacity to accurately assess its own performance, which can be limited by factors such as initial accuracy, question difficulty, and the lack of external feedback. In this paper, we delve into a two-player paradigm that separates the roles of reasoning and critique models, where the critique model provides step-level feedback to supervise the reasoning (actor) model during both test-time and train-time. We first propose AutoMathCritique, an automated and scalable framework for collecting critique data, resulting in a dataset of $76,321$ responses paired with step-level feedback. Fine-tuning language models with this dataset enables them to generate natural language feedback for mathematical reasoning. We demonstrate that the critique models consistently improve the actor's performance on difficult queries at test-time, especially when scaling up inference-time computation. Motivated by these findings, we introduce the critique-based supervision to the actor's self-training process, and propose a critique-in-the-loop self-improvement method. Experiments show that the method improves the actor's exploration efficiency and solution diversity, especially on challenging queries, leading to a stronger reasoning model. Lastly, we take the preliminary step to explore training self-talk reasoning models via critique supervision and showcase its potential. Our code and datasets are at \href{https://mathcritique.github.io/}{https://mathcritique.github.io/}.

cs.CL↗
arXiv2024

RoTBench: A Multi-Level Benchmark for Evaluating the Robustness of Large Language Models in Tool Learning

Tool learning has generated widespread interest as a vital means of interaction between Large Language Models (LLMs) and the physical world. Current research predominantly emphasizes LLMs' capacity to utilize tools in well-structured environments while overlooking their stability when confronted with the inevitable noise of the real world. To bridge this gap, we introduce RoTBench, a multi-level benchmark for evaluating the robustness of LLMs in tool learning. Specifically, we establish five external environments, each featuring varying levels of noise (i.e., Clean, Slight, Medium, Heavy, and Union), providing an in-depth analysis of the model's resilience across three critical phases: tool selection, parameter identification, and content filling. Experiments involving six widely-used models underscore the urgent necessity for enhancing the robustness of LLMs in tool learning. For instance, the performance of GPT-4 even drops significantly from 80.00 to 58.10 when there is no substantial change in manual accuracy. More surprisingly, the noise correction capability inherent in the GPT family paradoxically impedes its adaptability in the face of mild noise. In light of these findings, we propose RoTTuning, a strategy that enriches the diversity of training environments to bolster the robustness of LLMs in tool learning. The code and data are available at https://github.com/Junjie-Ye/RoTBench.

cs.CL↗
arXiv2024

ToolSword: Unveiling Safety Issues of Large Language Models in Tool Learning Across Three Stages

Tool learning is widely acknowledged as a foundational approach or deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present *ToolSword*, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing **malicious queries** and **jailbreak attacks** in the input stage, **noisy misdirection** and **risky cues** in the execution stage, and **harmful feedback** and **error conflicts** in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data is released in https://github.com/Junjie-Ye/ToolSword.

cs.CL↗
arXiv2024

Commuting Schemes of Upper Triangular Matrices and Representation Homology

We establish a connection between generalised commuting schemes $C_g(U_n)$ of higher genus $g$, which are associated with a group scheme $U_n$ consisting of upper triangular unipotent matrices, and the representation homology $HR_*(\Sigma_g,U_n)$ of a Riemann surface $\Sigma_g$ with coefficients in the group $U_n$. As an outcome, we provide a numerical criterion for determining whether commuting schemes $C(U_n)$ form a complete intersection.

math.AG↗
arXiv2024

ToolEyes: Fine-Grained Evaluation for Tool Learning Capabilities of Large Language Models in Real-world Scenarios

Existing evaluations of tool learning primarily focus on validating the alignment of selected tools for large language models (LLMs) with expected outcomes. However, these approaches rely on a limited set of scenarios where answers can be pre-determined, diverging from genuine needs. Furthermore, a sole emphasis on outcomes disregards the complex capabilities required for LLMs to effectively use tools. To tackle this issue, we propose ToolEyes, a fine-grained system tailored for the evaluation of the LLMs' tool learning capabilities in authentic scenarios. The system meticulously examines seven real-world scenarios, analyzing five dimensions crucial to LLMs in tool learning: format alignment, intent comprehension, behavior planning, tool selection, and answer organization. Additionally, ToolEyes incorporates a tool library boasting approximately 600 tools, serving as an intermediary between LLMs and the physical world. Evaluations involving ten LLMs across three categories reveal a preference for specific scenarios and limited cognitive abilities in tool learning. Intriguingly, expanding the model size even exacerbates the hindrance to tool learning. The code and data are available at https://github.com/Junjie-Ye/ToolEyes.

cs.CL↗
arXiv2023

TVDO: Tchebycheff Value-Decomposition Optimization for Multi-Agent Reinforcement Learning

In cooperative multiagent reinforcement learning (MARL), centralized training with decentralized execution (CTDE) has recently attracted more attention due to the physical demand. However, the most dilemma therein is the inconsistency between jointly-trained policies and individually-executed actions. In this article, we propose a factorized Tchebycheff value-decomposition optimization (TVDO) method to overcome the trouble of inconsistency. In particular, a nonlinear Tchebycheff aggregation function is formulated to realize the global optimum by tightly constraining the upper bound of individual action-value bias, which is inspired by the Tchebycheff method of multi-objective optimization. We theoretically prove that, under no extra limitations, the factorized value decomposition with Tchebycheff aggregation satisfies the sufficiency and necessity of Individual-Global-Max (IGM), which guarantees the consistency between the global and individual optimal action-value function. Empirically, in the climb and penalty game, we verify that TVDO precisely expresses the global-to-individual value decomposition with a guarantee of policy consistency. Meanwhile, we evaluate TVDO in the SMAC benchmark, and extensive experiments demonstrate that TVDO achieves a significant performance superiority over some SOTA MARL baselines.

cs.MA↗
arXiv2022

Temporal signals drive the emergence of multicellular information networks

Coordinated responses to environmental stimuli are critical for multicellular organisms. To overcome the obstacles of cell-to-cell heterogeneity and noisy signaling dynamics within individual cells, cells must effectively exchange information with peers. However, the dynamics and mechanisms of collective information transfer driven by external signals is poorly understood. Here we investigate the calcium dynamics of neuronal cells that form confluent monolayers and respond to cyclic ATP stimuli in microfluidic devices. Using Granger inference to reconstruct the underlying causal relations between the cells, we find that the cells self-organize into spatially decentralized and temporally stationary networks to support information transfer via gap junction channels. The connectivity of the causal networks depend on the temporal profile of the external stimuli, where short periods, or long periods with small duty fractions, lead to reduced connectivity and fractured network topology. We build a theoretical model based on communicating excitable units that reproduces our observations. The model further predicts that connectivity of the causal network is maximal at an optimal communication strength, which is confirmed by the experiments. Together, our results show that information transfer between neuronal cells is externally regulated by the temporal profile of the stimuli, and internally regulated by cell-cell communication.

physics.bio-ph↗
arXiv2021

An Unsupervised Domain Adaptation Model based on Dual-module Adversarial Training

In this paper, we propose a dual-module network architecture that employs a domain discriminative feature module to encourage the domain invariant feature module to learn more domain invariant features. The proposed architecture can be applied to any model that utilizes domain invariant features for unsupervised domain adaptation to improve its ability to extract domain invariant features. We conduct experiments with the Domain-Adversarial Training of Neural Networks (DANN) model as a representative algorithm. In the training process, we supply the same input to the two modules and then extract their feature distribution and prediction results respectively. We propose a discrepancy loss to find the discrepancy of the prediction results and the feature distribution between the two modules. Through the adversarial training by maximizing the loss of their feature distribution and minimizing the discrepancy of their prediction results, the two modules are encouraged to learn more domain discriminative and domain invariant features respectively. Extensive comparative evaluations are conducted and the proposed approach outperforms the state-of-the-art in most unsupervised domain adaptation tasks.

cs.LG↗
arXiv2021

An MAP Estimation for Between-Class Variance

Probabilistic linear discriminant analysis (PLDA) has been widely used in open-set verification tasks, such as speaker verification. A potential issue of this model is that the training set often contains limited number of classes, which makes the estimation for the between-class variance unreliable. This unreliable estimation often leads to degraded generalization. In this paper, we present an MAP estimation for the between-class variance, by employing an Inverse-Wishart prior. A key problem is that with hierarchical models such as PLDA, the prior is placed on the variance of class means while the likelihood is based on class members, which makes the posterior inference intractable. We derive a simple MAP estimation for such a model, and test it in both PLDA scoring and length normalization. In both cases, the MAP-based estimation delivers interesting performance improvement.

cs.SD↗