arXiv subjects

Qiyuan Zhang

Publications and source records attributed to Qiyuan Zhang.

28 records · Page 2Linked to original sources

arXiv2025

Crowd Comparative Reasoning: Unlocking Comprehensive Evaluations for LLM-as-a-Judge

LLM-as-a-Judge, which generates chain-of-thought (CoT) judgments, has become a widely adopted auto-evaluation method. However, its reliability is compromised by the CoT reasoning's inability to capture comprehensive and deeper details, often leading to incomplete outcomes. Existing methods mainly rely on majority voting or criteria expansion, which is insufficient to address the limitation in CoT. We propose Crowd-based Comparative Evaluation, which introduces additional crowd responses to compare with the candidate responses, thereby exposing deeper and more comprehensive details within the candidate responses. This process effectively guides LLM-as-a-Judge to provide a more detailed CoT judgment. Extensive experiments demonstrate that our approach enhances evaluation reliability, achieving an average accuracy gain of 6.7% across five benchmarks. Moreover, our method produces higher-quality CoTs that facilitate judge distillation and exhibit superior performance in rejection sampling for supervised fine-tuning (SFT), referred to as crowd rejection sampling, thereby enabling more efficient SFT. Our analysis confirms that CoTs generated by ours are more comprehensive and of higher quality, and evaluation accuracy improves as inference scales.

cs.CL↗

arXiv2025

HiLo: Learning Whole-Body Human-like Locomotion with Motion Tracking Controller

Deep Reinforcement Learning (RL) has emerged as a promising method to develop humanoid robot locomotion controllers. Despite the robust and stable locomotion demonstrated by previous RL controllers, their behavior often lacks the natural and agile motion patterns necessary for human-centric scenarios. In this work, we propose HiLo (human-like locomotion with motion tracking), an effective framework designed to learn RL policies that perform human-like locomotion. The primary challenges of human-like locomotion are complex reward engineering and domain randomization. HiLo overcomes these issues by developing an RL-based motion tracking controller and simple domain randomization through random force injection and action delay. Within the framework of HiLo, the whole-body control problem can be decomposed into two components: One part is solved using an open-loop control method, while the residual part is addressed with RL policies. A distributional value function is also implemented to stabilize the training process by improving the estimation of cumulative rewards under perturbed dynamics. Our experiments demonstrate that the motion tracking controller trained using HiLo can perform natural and agile human-like locomotion while exhibiting resilience to external disturbances in real-world systems. Furthermore, we show that the motion patterns of humanoid robots can be adapted through the residual mechanism without fine-tuning, allowing quick adjustments to task requirements.

cs.RO↗

arXiv2024

UniAvatar: Taming Lifelike Audio-Driven Talking Head Generation with Comprehensive Motion and Lighting Control

Recently, animating portrait images using audio input is a popular task. Creating lifelike talking head videos requires flexible and natural movements, including facial and head dynamics, camera motion, realistic light and shadow effects. Existing methods struggle to offer comprehensive, multifaceted control over these aspects. In this work, we introduce UniAvatar, a designed method that provides extensive control over a wide range of motion and illumination conditions. Specifically, we use the FLAME model to render all motion information onto a single image, maintaining the integrity of 3D motion details while enabling fine-grained, pixel-level control. Beyond motion, this approach also allows for comprehensive global illumination control. We design independent modules to manage both 3D motion and illumination, permitting separate and combined control. Extensive experiments demonstrate that our method outperforms others in both broad-range motion control and lighting control. Additionally, to enhance the diversity of motion and environmental contexts in current datasets, we collect and plan to publicly release two datasets, DH-FaceDrasMvVid-100 and DH-FaceReliVid-200, which capture significant head movements during speech and various lighting scenarios.

cs.CV↗

arXiv2024

Collaborative Performance Prediction for Large Language Models

Comprehensively understanding and accurately predicting the performance of large language models across diverse downstream tasks has emerged as a pivotal challenge in NLP research. The pioneering scaling law on downstream works demonstrated intrinsic similarities within model families and utilized such similarities for performance prediction. However, they tend to overlook the similarities between model families and only consider design factors listed in the original scaling law. To overcome these limitations, we introduce a novel framework, Collaborative Performance Prediction (CPP), which significantly enhances prediction accuracy by leveraging the historical performance of various models on downstream tasks and other design factors for both model and task. We also collect a collaborative data sourced from online platforms containing both historical performance and additional design factors. With the support of the collaborative data, CPP not only surpasses traditional scaling laws in predicting the performance of scaled LLMs but also facilitates a detailed analysis of factor importance, an area previously overlooked.

cs.CL↗

arXiv2024

Heat conduction in low-dimensional electron gases without and with a magnetic field

We investigate the behavior of heat conduction in two-dimensional (2D) electron gases without and with a magnetic field. We perform simulations with the Multi-Particle-Collision approach, suitably adapted to account for the Lorenz force acting on the particles. For zero magnetic field, we find that the heat conductivity $κ$ diverges with the system size $L$ following the logarithmic relation $κ\thicksim \ln L$ (as predicted for two-dimensional (2D) systems) for small $L$ values; however, in the thermodynamic limit the heat conductivity tends to follow the relation $κ\thicksim L^{1/3}$, as predicted for one-dimensional (1D) fluids. This suggests the presence of a dimensional-crossover effect in heat conduction in electronic systems that adhere to standard momentum conservation. Under the magnetic field, time-reversal symmetry is broken and the standard momentum conservation in the system is no longer satisfied but the \emph{pseudomomentum} of the system is still conserved. In contrast with the zero-field case, both equilibrium and non-equilibrium simulations indicate a finite heat conductivity independent on the system size $L$ as $L$ increases. This indicates that pseudomomentum conservation can exhibit normal diffusive heat transport, which differs from the abnormal behavior observed in low-dimensional coupled charged harmonic oscillators with pseudomomentum conservation in a magnetic field. These findings support the validity of the hydrodynamic theory in electron gases and clarify that pseudomomentum conservation is not enough to ensure the anomalous behavior of heat conduction.

cond-mat.mes-hall↗

arXiv2022

Introducing RISK

This extended abstract introduces the initial steps taken to develop a system for Rapid Internal Simulation of Knowledge (RISK). RISK aims to enable more transparency in artificial intelligence systems, especially those created by deep learning networks by allowing real-time simulation of what the system knows. By looking at hypothetical situations based on these simulations a system may make more informed decisions, and produce them for non-expert observers to understand the reasoning behind a given action.

cs.RO↗

arXiv2021

Believe What You See: Implicit Constraint Approach for Offline Multi-Agent Reinforcement Learning

Learning from datasets without interaction with environments (Offline Learning) is an essential step to apply Reinforcement Learning (RL) algorithms in real-world scenarios. However, compared with the single-agent counterpart, offline multi-agent RL introduces more agents with the larger state and action space, which is more challenging but attracts little attention. We demonstrate current offline RL algorithms are ineffective in multi-agent systems due to the accumulated extrapolation error. In this paper, we propose a novel offline RL algorithm, named Implicit Constraint Q-learning (ICQ), which effectively alleviates the extrapolation error by only trusting the state-action pairs given in the dataset for value estimation. Moreover, we extend ICQ to multi-agent tasks by decomposing the joint-policy under the implicit constraint. Experimental results demonstrate that the extrapolation error is successfully controlled within a reasonable range and insensitive to the number of agents. We further show that ICQ achieves the state-of-the-art performance in the challenging multi-agent offline tasks (StarCraft II). Our code is public online at https://github.com/YiqinYang/ICQ.

cs.AI↗

arXiv2021

NOAHQA: Numerical Reasoning with Interpretable Graph Question Answering Dataset

While diverse question answering (QA) datasets have been proposed and contributed significantly to the development of deep learning models for QA tasks, the existing datasets fall short in two aspects. First, we lack QA datasets covering complex questions that involve answers as well as the reasoning processes to get the answers. As a result, the state-of-the-art QA research on numerical reasoning still focuses on simple calculations and does not provide the mathematical expressions or evidences justifying the answers. Second, the QA community has contributed much effort to improving the interpretability of QA models. However, these models fail to explicitly show the reasoning process, such as the evidence order for reasoning and the interactions between different pieces of evidence. To address the above shortcomings, we introduce NOAHQA, a conversational and bilingual QA dataset with questions requiring numerical reasoning with compound mathematical expressions. With NOAHQA, we develop an interpretable reasoning graph as well as the appropriate evaluation metric to measure the answer quality. We evaluate the state-of-the-art QA models trained using existing QA datasets on NOAHQA and show that the best among them can only achieve 55.5 exact match scores, while the human performance is 89.7. We also present a new QA model for generating a reasoning graph where the reasoning graph metric still has a large gap compared with that of humans, e.g., 28 scores.

cs.CL↗

arXiv2021

MWPToolkit: An Open-Source Framework for Deep Learning-Based Math Word Problem Solvers

Developing automatic Math Word Problem (MWP) solvers has been an interest of NLP researchers since the 1960s. Over the last few years, there are a growing number of datasets and deep learning-based methods proposed for effectively solving MWPs. However, most existing methods are benchmarked soly on one or two datasets, varying in different configurations, which leads to a lack of unified, standardized, fair, and comprehensive comparison between methods. This paper presents MWPToolkit, the first open-source framework for solving MWPs. In MWPToolkit, we decompose the procedure of existing MWP solvers into multiple core components and decouple their models into highly reusable modules. We also provide a hyper-parameter search function to boost the performance. In total, we implement and compare 17 MWP solvers on 4 widely-used single equation generation benchmarks and 2 multiple equations generation benchmarks. These features enable our MWPToolkit to be suitable for researchers to reproduce advanced baseline models and develop new MWP solvers quickly. Code and documents are available at https://github.com/LYH-YF/MWPToolkit.

cs.CL↗

arXiv2021

Estimation of transmitted wavefronts at defocused positions in a broad bandwidth range

Wavefront aberrations can reflect the imaging quality of high-performance optical systems better than geometric aberrations. Although laser interferometers have emerged as the main tool for measurement of transmitted wavefronts, their application is greatly limited, as they are typically designed for operation at specific wavelengths. In a previous study, we proposed a method for determining the wavefront transmitted by an optical system at any wavelength in a certain band. Although this method works well for most monochromatic systems, where the image plane is at the focal point for the transmission wavelength, for general multi-color systems, it is more practical to measure the wavefront at the defocused image plane. Hence, in this paper, we have developed a complete method for determining transmitted wavefronts in a broad bandwidth at any defocused position, enabling wavefront measurements for multi-color systems. Here, we assume that in small ranges, the Zernike coefficients have a linear relationship with position, such that Zernike coefficients at defocused positions can be derived from measurements performed at the focal point. We conducted experiments to verify these assumptions, validating the new method. The experimental setup has been improved so that it can handle multi-color systems, and a detailed experimental process is summarized. With this technique, application of broadband transmission wavefront measurement can be extended to most general optical systems, which is of great significance for characterization of achromatic and apochromatic optical lenses.

physics.optics↗