arXiv subjects

Yilong Wu

Publications and source records attributed to Yilong Wu.

15 recordsLinked to original sources

arXiv2026

Generalizations of the M&M Game

The MandM Game involves two players who begin with I1 and I2 MandM's. During each round, each player tosses a fair coin: if the coin lands heads, that player eats one MandM, and if it lands tails, the player does not eat. If, at the end of a round, one player still has MandM's while the other has none, then the player with MandM's remaining is declared the winner. If both players eat their last MandM in the same round, the game is said to end in a tie. In [BHM+17], the authors studied the probability of a tie in the MandM Game and derived a simple closed-form expression in the special case where both players start with the same number of MandM's. We generalize the MandM Game in several directions, including allowing players to toss multiple coins per round and modifying the probability distributions of the coin flips. We use the technique of generating functions, Monte Carlo methods, and non-linear curve fitting to study the generalized MandM Game.

math.PR↗

arXiv2026

MulDimIF: A Multi-Dimensional Constraint Framework for Evaluating and Improving Instruction Following in Large Language Models

Instruction following refers to the ability of large language models (LLMs) to generate outputs that satisfy all specified constraints. Existing research has primarily focused on constraint categories, offering limited evaluation dimensions and little guidance for improving instruction-following abilities. To address this gap, we introduce MulDimIF, a multi-dimensional constraint framework encompassing three constraint patterns, four constraint categories, and four difficulty levels. Based on this framework, we design a controllable instruction generation pipeline. Through constraint expansion, conflict detection, and instruction rewriting, we construct 9,106 code-verifiable samples. We evaluate 18 LLMs from six model families and find marked performance differences across constraint settings. For instance, average accuracy decreases from 80.82% at Level I to 36.76% at Level IV. Moreover, training with data generated by our framework significantly improves instruction following without compromising general performance. In-depth analysis indicates that these gains stem largely from parameter updates in attention modules, which strengthen constraint recognition and adherence. Code and data are available in https://github.com/Junjie-Ye/MulDimIF.

cs.CL↗

arXiv2026

LLMEval-Fair: A Large-Scale Longitudinal Study on Robust and Fair Evaluation of Large Language Models

Existing evaluation of Large Language Models (LLMs) on static benchmarks is vulnerable to data contamination and leaderboard overfitting, critical issues that obscure true model capabilities. To address this, we introduce LLMEval-Fair, a framework for dynamic evaluation of LLMs. LLMEval-Fair is built on a proprietary bank of 220k graduate-level questions, from which it dynamically samples unseen test sets for each evaluation run. Its automated pipeline ensures integrity via contamination-resistant data curation, a novel anti-cheating architecture, and a calibrated LLM-as-a-judge process achieving 90% agreement with human experts, complemented by a relative ranking system for fair comparison. A 30-month longitudinal study of nearly 60 leading models reveals a performance ceiling on knowledge memorization and exposes data contamination vulnerabilities undetectable by static benchmarks. The framework demonstrates exceptional robustness in ranking stability and consistency, providing strong empirical validation for the dynamic evaluation paradigm. LLMEval-Fair offers a robust and credible methodology for assessing the true capabilities of LLMs beyond leaderboard scores, promoting the development of more trustworthy evaluation standards. Our code and data are publicly available at https://github.com/llmeval/LLMEval-Fair.

cs.CL↗

arXiv2026

SpeechRole: A Large-Scale Dataset and Benchmark for Evaluating Speech Role-Playing Agents

Speech is essential for realistic role-playing, yet existing work on role-playing agents largely centers on text, leaving Speech Role-Playing Agents (SRPAs) underexplored and without systematic evaluation. We introduce SpeechRole, a unified framework for developing and assessing SRPAs. SpeechRole-Data contains 98 roles and 111k speech-to-speech conversations with rich timbre and prosodic variation, providing large-scale resources for training SRPAs. SpeechRole-Eval offers a multidimensional benchmark that directly evaluates generated speech, preserving paralinguistic cues and measuring interaction ability, speech expressiveness, and role-playing fidelity. Experiments show that end-to-end SRPAs such as GPT-4o Audio achieve strong fluency and naturalness, but remain limited in prosody consistency and emotion appropriateness. In contrast, current open-source end-to-end models exhibit substantial performance gaps across multiple evaluation dimensions. Cascaded and end-to-end systems achieve comparable results in interaction ability and role-playing fidelity, suggesting that these aspects are still largely influenced by the underlying text-based language models. We release all data, code, and evaluation tools at https://github.com/yuhui1038/SpeechRole.

cs.CL↗

arXiv2025

TL-Training: A Task-Feature-Based Framework for Training Large Language Models in Tool Use

Large language models (LLMs) achieve remarkable advancements by leveraging tools to interact with environments, a critical step toward generalized AI. However, the standard supervised fine-tuning (SFT) approach, which relies on large-scale datasets, often overlooks task-specific characteristics in tool use, leading to performance bottlenecks. To address this issue, we analyze three existing LLMs and uncover key insights: training data can inadvertently impede tool-use behavior, token importance is distributed unevenly, and errors in tool calls fall into a small set of categories. Building on these findings, we propose~\emph{TL-Training}, a task-feature-based framework that mitigates the effects of suboptimal training data, dynamically adjusts token weights to prioritize key tokens during SFT, and incorporates a robust reward mechanism tailored to error categories, optimized through proximal policy optimization. We validate TL-Training by training CodeLLaMA-2-7B and evaluating it on four open-source test sets. Our results demonstrate that the LLM trained by our method matches or surpasses both open- and closed-source LLMs in tool-use performance using only 1,217 training data points. Additionally, our method enhances robustness in noisy environments and improves general task performance, offering a scalable and efficient paradigm for tool-use training in LLMs. Code and data are available at https://github.com/Junjie-Ye/TL-Training.

cs.CL↗

arXiv2025

Feature-free regression kriging

Spatial interpolation is a crucial task in geography. As perhaps the most widely used interpolation methods, geostatistical models -- such as Ordinary Kriging (OK) -- assume spatial stationarity, which makes it difficult to capture the nonstationary characteristics of geographic variables. A common solution is trend surface modeling (e.g., Regression Kriging, RK), which relies on external explanatory variables to model the trend and then applies geostatistical interpolation to the residuals. However, this approach requires high-quality and readily available explanatory variables, which are often lacking in many spatial interpolation scenarios -- such as estimating heavy metal concentrations underground. This study proposes a Feature-Free Regression Kriging (FFRK) method, which automatically extracts geospatial features -- including local dependence, local heterogeneity, and geosimilarity -- to construct a regression-based trend surface without requiring external explanatory variables. We conducted experiments on the spatial distribution prediction of three heavy metals in a mining area in Australia. In comparison with 17 classical interpolation methods, the results indicate that FFRK, which does not incorporate any explanatory variables and relies solely on extracted geospatial features, consistently outperforms both conventional Kriging techniques and machine learning models that depend on explanatory variables. This approach effectively addresses spatial nonstationarity while reducing the cost of acquiring explanatory variables, improving both prediction accuracy and generalization ability. This finding suggests that an accurate characterization of geospatial features based on domain knowledge can significantly enhance spatial prediction performance -- potentially yielding greater improvements than merely adopting more advanced statistical models.

physics.soc-ph↗

arXiv2025

PFDial: A Structured Dialogue Instruction Fine-tuning Method Based on UML Flowcharts

Process-driven dialogue systems, which operate under strict predefined process constraints, are essential in customer service and equipment maintenance scenarios. Although Large Language Models (LLMs) have shown remarkable progress in dialogue and reasoning, they still struggle to solve these strictly constrained dialogue tasks. To address this challenge, we construct Process Flow Dialogue (PFDial) dataset, which contains 12,705 high-quality Chinese dialogue instructions derived from 440 flowcharts containing 5,055 process nodes. Based on PlantUML specification, each UML flowchart is converted into atomic dialogue units i.e., structured five-tuples. Experimental results demonstrate that a 7B model trained with merely 800 samples, and a 0.5B model trained on total data both can surpass 90% accuracy. Additionally, the 8B model can surpass GPT-4o up to 43.88% with an average of 11.00%. We further evaluate models' performance on challenging backward transitions in process flows and conduct an in-depth analysis of various dataset formats to reveal their impact on model performance in handling decision and sequential branches. The data is released in https://github.com/KongLongGeFDU/PFDial.

cs.CL↗

arXiv2025

STDArm: Transferring Visuomotor Policies From Static Data Training to Dynamic Robot Manipulation

Recent advances in mobile robotic platforms like quadruped robots and drones have spurred a demand for deploying visuomotor policies in increasingly dynamic environments. However, the collection of high-quality training data, the impact of platform motion and processing delays, and limited onboard computing resources pose significant barriers to existing solutions. In this work, we present STDArm, a system that directly transfers policies trained under static conditions to dynamic platforms without extensive modifications. The core of STDArm is a real-time action correction framework consisting of: (1) an action manager to boost control frequency and maintain temporal consistency, (2) a stabilizer with a lightweight prediction network to compensate for motion disturbances, and (3) an online latency estimation module for calibrating system parameters. In this way, STDArm achieves centimeter-level precision in mobile manipulation tasks. We conduct comprehensive evaluations of the proposed STDArm on two types of robotic arms, four types of mobile platforms, and three tasks. Experimental results indicate that the STDArm enables real-time compensation for platform motion disturbances while preserving the original policy's manipulation capabilities, achieving centimeter-level operational precision during robot motion.

cs.RO↗

arXiv2025

MT-PCR: Leveraging Modality Transformation for Large-Scale Point Cloud Registration with Limited Overlap

Large-scale scene point cloud registration with limited overlap is a challenging task due to computational load and constrained data acquisition. To tackle these issues, we propose a point cloud registration method, MT-PCR, based on Modality Transformation. MT-PCR leverages a BEV capturing the maximal overlap information to improve the accuracy and utilizes images to provide complementary spatial features. Specifically, MT-PCR converts 3D point clouds to BEV images and eastimates correspondence by 2D image keypoints extraction and matching. Subsequently, the 2D correspondence estimates are then transformed back to 3D point clouds using inverse mapping. We have applied MT-PCR to Terrestrial Laser Scanning and Aerial Laser Scanning point cloud registration on the GrAco dataset, involving 8 low-overlap, square-kilometer scale registration scenarios. Experiments and comparisons with commonly used methods demonstrate that MT-PCR can achieve superior accuracy and robustness in large-scale scenes with limited overlap.

cs.RO↗

arXiv2025

OG-Gaussian: Occupancy Based Street Gaussians for Autonomous Driving

Accurate and realistic 3D scene reconstruction enables the lifelike creation of autonomous driving simulation environments. With advancements in 3D Gaussian Splatting (3DGS), previous studies have applied it to reconstruct complex dynamic driving scenes. These methods typically require expensive LiDAR sensors and pre-annotated datasets of dynamic objects. To address these challenges, we propose OG-Gaussian, a novel approach that replaces LiDAR point clouds with Occupancy Grids (OGs) generated from surround-view camera images using Occupancy Prediction Network (ONet). Our method leverages the semantic information in OGs to separate dynamic vehicles from static street background, converting these grids into two distinct sets of initial point clouds for reconstructing both static and dynamic objects. Additionally, we estimate the trajectories and poses of dynamic objects through a learning-based approach, eliminating the need for complex manual annotations. Experiments on Waymo Open dataset demonstrate that OG-Gaussian is on par with the current state-of-the-art in terms of reconstruction quality and rendering speed, achieving an average PSNR of 35.13 and a rendering speed of 143 FPS, while significantly reducing computational costs and economic overhead.

cs.CV↗

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↗

arXiv2024

TransferTOD: A Generalizable Chinese Multi-Domain Task-Oriented Dialogue System with Transfer Capabilities

Task-oriented dialogue (TOD) systems aim to efficiently handle task-oriented conversations, including information collection. How to utilize TOD accurately, efficiently and effectively for information collection has always been a critical and challenging task. Recent studies have demonstrated that Large Language Models (LLMs) excel in dialogue, instruction generation, and reasoning, and can significantly enhance the performance of TOD through fine-tuning. However, current datasets primarily cater to user-led systems and are limited to predefined specific scenarios and slots, thereby necessitating improvements in the proactiveness, diversity, and capabilities of TOD. In this study, we present a detailed multi-domain task-oriented data construction process for conversations, and a Chinese dialogue dataset generated based on this process, TransferTOD, which authentically simulates human-computer dialogues in 30 popular life service scenarios. Leveraging this dataset, we trained a model called TransferTOD-7B using full-parameter fine-tuning, showcasing notable abilities in slot filling and questioning. Our work has demonstrated its strong generalization capabilities in various downstream scenarios, significantly enhancing both data utilization efficiency and system performance. The data is released in https://github.com/KongLongGeFDU/TransferTOD.

cs.AI↗

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

Rotation Initialization and Stepwise Refinement for Universal LiDAR Calibration

Autonomous systems often employ multiple LiDARs to leverage the integrated advantages, enhancing perception and robustness. The most critical prerequisite under this setting is the estimating the extrinsic between each LiDAR, i.e., calibration. Despite the exciting progress in multi-LiDAR calibration efforts, a universal, sensor-agnostic calibration method remains elusive. According to the coarse-to-fine framework, we first design a spherical descriptor TERRA for 3-DoF rotation initialization with no prior knowledge. To further optimize, we present JEEP for the joint estimation of extrinsic and pose, integrating geometric and motion information to overcome factors affecting the point cloud registration. Finally, the LiDAR poses optimized by the hierarchical optimization module are input to time synchronization module to produce the ultimate calibration results, including the time offset. To verify the effectiveness, we conduct extensive experiments on eight datasets, where 16 diverse types of LiDARs in total and dozens of calibration tasks are tested. In the challenging tasks, the calibration errors can still be controlled within 5cm and 1° with a high success rate.

cs.RO↗