arXiv subjects

Shijie Han

Publications and source records attributed to Shijie Han.

9 recordsLinked to original sources

arXiv2026

Coalition-Aware Skill Reliability for Self-Evolving Agents

Agent skills, structured artifacts distilled from interaction trajectories and dynamically reused from skill banks, have become a central mechanism for enabling large language model (LLM)-based self-evolving agents to learn from past experience. Yet existing work has largely focused on the operational aspects of skills, such as acquisition, evolution, and retrieval, while leaving a more fundamental reliability question unresolved: Do accumulated skills in an agent's skill bank actually make positive mechanistic contributions? We investigate this question through systematic skill-bank audits across alternative bank compositions and deployment domains, measuring the resulting changes in agent behavior. These audits reveal two recurring reliability failures: coalition pollution, where bank-level gains conceal negative coalition-level skill contributions, and cross-domain utility reversal, where source-beneficial skills reverse their effects after transfer. These findings motivate two reliability interventions: coalition-aware skill selection during skill accumulation and label-free skill masking after transfer. Coalition-Aware Skill Selection (CASS) selects more reliable candidate skills for the current bank using sampled Shapley marginals. Unsupervised Skill-Masked Coalition Optimizer (u-SMCO) masks transferred skills whose exclusion improves retrieval quality on unlabeled target-domain data. Agentic experiments on LoCoMo, LongMemEval, HotpotQA, and ALFWorld show that CASS and u-SMCO consistently improve task performance and cross-domain generalization over strong skill-based self-evolving agent baselines. Beyond accuracy, coalition-conditioned reliability modeling reduces sensitivity to noisy outcome-reward fluctuations during reinforcement learning and exposes the limits of isolation-based skill evaluation.

cs.AI↗

arXiv2025

High-Precision Temperature Estimation Based on Magnetic Nanoparticles Dominated by Brownian Relaxation under Combined AC and DC Magnetic Fields

Brownian relaxation is one of the primary mechanisms that allows magnetic nanoparticles (MNPs) to convert magnetic energy into thermal energy under an excitation magnetic field. Accurately characterizing the MNPs' magnetization dynamics dominated by Brownian relaxation is crucial for achieving high-precision temperature estimation. However, the lack of a readily applicable analytical expression remains a major obstacle to the advancement of magnetic nanoparticle hyperthermia (MNPH). In this paper, the perturbation method was applied to derive analytical expressions from the Fokker-Planck equation, which characterized MNPs' magnetization behaviors under the AC and DC magnetic fields. Numerical simulations were conducted to validate the accuracy of the analytical expressions and to explore the correlation between temperature and the magnetization response. Then, a temperature analysis model based on magnetization harmonics was constructed. The first and second harmonic ratios and first harmonic phase were used to calculate MNPs' temperature, respectively. The experimental results demonstrate that within 310 K to 320 K, the estimation error of the temperature using the amplitude ratio of the first to second harmonics is below 0.0151 K, while the error using the first harmonic phase is below 0.0218 K. The derived analytical expressions are expected to enhance the accuracy of MNP-based temperature measurements and facilitate their broader applications in MNPH and MNP imaging.

cond-mat.mes-hall↗

arXiv2025

FinSphere, a Real-Time Stock Analysis Agent Powered by Instruction-Tuned LLMs and Domain Tools

Current financial large language models (FinLLMs) struggle with two critical limitations: the absence of objective evaluation metrics to assess the quality of stock analysis reports and a lack of depth in stock analysis, which impedes their ability to generate professional-grade insights. To address these challenges, this paper introduces FinSphere, a stock analysis agent, along with three major contributions: (1) AnalyScore, a systematic evaluation framework for assessing stock analysis quality, (2) Stocksis, a dataset curated by industry experts to enhance LLMs' stock analysis capabilities, and (3) FinSphere, an AI agent that can generate high-quality stock analysis reports in response to user queries. Experiments demonstrate that FinSphere achieves superior performance compared to both general and domain-specific LLMs, as well as existing agent-based systems, even when they are enhanced with real-time data access and few-shot guidance. The integrated framework, which combines real-time data feeds, quantitative tools, and an instruction-tuned LLM, yields substantial improvements in both analytical quality and practical applicability for real-world stock analysis.

cs.AI↗

arXiv2024

QPruner: Probabilistic Decision Quantization for Structured Pruning in Large Language Models

The rise of large language models (LLMs) has significantly advanced various natural language processing (NLP) tasks. However, the resource demands of these models pose substantial challenges. Structured pruning is an effective approach to reducing model size, but it often results in significant accuracy degradation, necessitating parameter updates to adapt. Unfortunately, such fine-tuning requires substantial memory, which limits its applicability. To address these challenges, we introduce quantization into the structured pruning framework to reduce memory consumption during both fine-tuning and inference. However, the combined errors from pruning and quantization increase the difficulty of fine-tuning, requiring a more refined quantization scheme. To this end, we propose QPruner, a novel framework that employs structured pruning to reduce model size, followed by a layer-wise mixed-precision quantization scheme. Quantization precisions are assigned to each layer based on their importance to the target task, and Bayesian optimization is employed to refine precision allocation strategies, ensuring a balance between model accuracy and memory efficiency. Extensive experiments on benchmark datasets demonstrate that QPruner significantly outperforms existing methods in memory savings while maintaining or improving model performance.

cs.LG↗

arXiv2024

RankAdaptor: Hierarchical Rank Allocation for Efficient Fine-Tuning Pruned LLMs via Performance Model

The efficient compression of large language models (LLMs) has become increasingly popular. However, recovering the performance of compressed LLMs remains a major challenge. The current practice in LLM compression entails the implementation of structural pruning, complemented by a recovery phase that leverages the Low-Rank Adaptation (LoRA) algorithm. Structural pruning's uneven modification of model architecture, coupled with standard LoRA's fixed configuration allocation across layers in an online pipeline, leads to suboptimal performance in various downstream tasks for pruned models. To address this challenge, we introduce RankAdaptor, a hierarchical rank allocation method that enables efficient fine-tuning of pruned LLMs according to layerwise specific recovery requirements. We employ a performance model that conducts offline meta-learning and online incremental learning to explore optimal rank values for each layer. Comprehensive experiments on popular benchmarks show that RankAdaptor consistently outperforms state-of-the-art methods across a variety of pruning settings and LLM architectures, with improvements ranging from 0.7\% to 5.5\%.

cs.CL↗

arXiv2024

Healing Powers of BERT: How Task-Specific Fine-Tuning Recovers Corrupted Language Models

Language models like BERT excel at sentence classification tasks due to extensive pre-training on general data, but their robustness to parameter corruption is unexplored. To understand this better, we look at what happens if a language model is "broken", in the sense that some of its parameters are corrupted and then recovered by fine-tuning. Strategically corrupting BERT variants at different levels, we find corrupted models struggle to fully recover their original performance, with higher corruption causing more severe degradation. Notably, bottom-layer corruption affecting fundamental linguistic features is more detrimental than top-layer corruption. Our insights contribute to understanding language model robustness and adaptability under adverse conditions, informing strategies for developing resilient NLP systems against parameter perturbations.

cs.CL↗

arXiv2024

Auxiliary Reward Generation with Transition Distance Representation Learning

Reinforcement learning (RL) has shown its strength in challenging sequential decision-making problems. The reward function in RL is crucial to the learning performance, as it serves as a measure of the task completion degree. In real-world problems, the rewards are predominantly human-designed, which requires laborious tuning, and is easily affected by human cognitive biases. To achieve automatic auxiliary reward generation, we propose a novel representation learning approach that can measure the ``transition distance'' between states. Building upon these representations, we introduce an auxiliary reward generation technique for both single-task and skill-chaining scenarios without the need for human knowledge. The proposed approach is evaluated in a wide range of manipulation tasks. The experiment results demonstrate the effectiveness of measuring the transition distance between states and the induced improvement by auxiliary rewards, which not only promotes better learning efficiency but also increases convergent stability.

cs.LG↗

arXiv2023

A Novel Estimation Method for Temperature of Magnetic Nanoparticles Dominated by Brownian Relaxation Based on Magnetic Particle Spectroscopy

This paper presents a novel method for estimating the temperature of magnetic nanoparticles (MNPs) based on AC magnetization harmonics of MNPs dominated by Brownian relaxation. The difference in the AC magnetization response and magnetization harmonic between the Fokker-Planck equation and the Langevin function was analyzed, and we studied the relationship between the magnetization harmonic and the key factors, such as Brownian relaxation time, temperature, magnetic field strength, core size and hydrodynamic size of MNPs, excitation frequency, and so on. We proposed a compensation function for AC magnetization harmonic with consideration of the key factors and the difference between the Fokker-Planck equation and the Langevin function. Then a temperature estimation model based on the compensation function and the Langevin function was established. By employing the least squares algorithm, the temperature was successfully calculated. The experimental results show that the temperature error is less than 0.035 K in the temperature range from 310 K to 320 K. The temperature estimation model is expected to improve the performance of the magnetic nanoparticle thermometer and be applied to magnetic nanoparticle-mediated hyperthermia.

physics.med-ph↗

arXiv2022

Classifying Ambiguous Identities in Hidden-Role Stochastic Games with Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning (MARL) is a prevalent learning paradigm for solving stochastic games. In most MARL studies, agents in a game are defined as teammates or enemies beforehand, and the relationships among the agents remain fixed throughout the game. However, in real-world problems, the agent relationships are commonly unknown in advance or dynamically changing. Many multi-party interactions start off by asking: who is on my team? This question arises whether it is the first day at the stock exchange or the kindergarten. Therefore, training policies for such situations in the face of imperfect information and ambiguous identities is an important problem that needs to be addressed. In this work, we develop a novel identity detection reinforcement learning (IDRL) framework that allows an agent to dynamically infer the identities of nearby agents and select an appropriate policy to accomplish the task. In the IDRL framework, a relation network is constructed to deduce the identities of other agents by observing the behaviors of the agents. A danger network is optimized to estimate the risk of false-positive identifications. Beyond that, we propose an intrinsic reward that balances the need to maximize external rewards and accurate identification. After identifying the cooperation-competition pattern among the agents, IDRL applies one of the off-the-shelf MARL methods to learn the policy. To evaluate the proposed method, we conduct experiments on Red-10 card-shedding game, and the results show that IDRL achieves superior performance over other state-of-the-art MARL methods. Impressively, the relation network has the par performance to identify the identities of agents with top human players; the danger network reasonably avoids the risk of imperfect identification. The code to reproduce all the reported results is available online at https://github.com/MR-BENjie/IDRL.

cs.AI↗