Yuchen Fan

PI-Mem: Pushing Long-Context Reasoning to 3.6M Tokens with Parallel-Iterative Memory

Dawei Liu, Haixu Song, Shuang Cheng, Shijie Wang et autres

Long-context reasoning remains a critical bottleneck for large language models, as recent recurrent-memory approaches face two inherent challenges: sequential chunk-wise updates can overwrite early critical evidence with later irrelevant content, and serial inter-chunk dependencies limit parallelism and cause latency to increase with …

cn, ca (code pays fourni par la source)

EVA-Score: Evaluating Abstractive Long-Form Summarization on Informativeness through Extraction and Validation

Yuchen Fan, Yazhe Wan, Xin Zhong, Haonan Cheng et autres

Recent advances in large language models (LLMs) have renewed interest in abstractive long-form summarization, where lengthy inputs carry high information density. Yet automatic evaluation remains inadequate: token-overlap metrics (ROUGE, BERTScore) favor surface similarity, and LLM-based judges degrade with extended context. We introduce …

cn (code pays fourni par la source)

FFV-PINN: A Fast Physics-Informed Neural Network with Simplified Finite Volume Discretization and Residual Correction

Chang Wei, Yuchen Fan, Jian Cheng Wong, Chin Chun Ooi et autres

Physics-informed neural networks (PINNs) have emerged as a major research focus. However, today's PINNs encounter several limitations. Firstly, during the construction of the loss function using automatic differentiation, PINNs often neglect information from neighboring points, which hinders their ability to enforce physical …

Scale-PINN: Learning Efficient Physics-Informed Neural Networks Through Sequential Correction

Pao-Hsiung Chiu, Jian Cheng Wong, Chin Chun Ooi, Chang Wei et autres

Physics-informed neural networks (PINNs) have emerged as a promising mesh-free paradigm for solving partial differential equations, yet adoption in science and engineering is limited by slow training and modest accuracy relative to modern numerical solvers. We introduce the Sequential Correction Algorithm for …

sg, cn (code pays fourni par la source)

Scale-PINN: Learning Efficient Physics-Informed Neural Networks Through Sequential Correction

Pao-Hsiung Chiu, Jian Cheng Wong, Chin Chun Ooi, Chang Wei et autres

Physics-informed neural networks (PINNs) have emerged as a promising mesh-free paradigm for solving partial differential equations, yet adoption in science and engineering is limited by slow training and modest accuracy relative to modern numerical solvers. We introduce the Sequential Correction Algorithm for …

Preference-Oriented Supervised Fine-Tuning: Favoring Target Model over Aligned Large Language Models

Yuchen Fan, Yuzhong Hong, Qiushi Wang, Junwei Bao et autres

Alignment, endowing a pre-trained Large language model (LLM) with the ability to follow instructions, is crucial for its real-world applications. Conventional supervised fine-tuning (SFT) methods formalize it as causal language modeling typically with a cross-entropy objective, requiring a large amount of high-quality …

Physics-Informed Neural Network Based on Control Volumes for Solving Time-Independent Problems

Yuchen Fan, Yongqing Zhou, Xin Liu, Chi Li et autres

cn (code pays fourni par la source)

Preference-Oriented Supervised Fine-Tuning: Favoring Target Model Over Aligned Large Language Models

Yuchen Fan, Yuzhong Hong, Qiushi Wang, Junwei Bao et autres

Alignment, endowing a pre-trained Large language model (LLM) with the ability to follow instructions, is crucial for its real-world applications. Conventional supervised fine-tuning (SFT) methods formalize it as causal language modeling typically with a cross-entropy objective, requiring a large amount of high-quality …

BoRA: Bi-dimensional Weight-Decomposed Low-Rank Adaptation

Qiushi Wang, Yuchen Fan, Junwei Bao, Yang Song

In recent years, Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) have significantly enhanced the adaptability of large-scale pre-trained models. Weight-Decomposed Low-Rank Adaptation (DoRA) improves upon LoRA by separating the magnitude and direction components of the weight matrix, leading to superior performance. …

Evaluating Generative Language Models in Information Extraction as Subjective Question Correction

Yuchen Fan, Yantao Liu, Zijun Yao, Jifan Yu et autres

Untangle the KNOT: Interweaving Conflicting Knowledge and Reasoning Skills in Large Language Models

Yantao Liu, Zijun Yao, Xin Lv, Yuchen Fan et autres

Providing knowledge documents for large language models (LLMs) has emerged as a promising solution to update the static knowledge inherent in their parameters. However, knowledge in the document may conflict with the memory of LLMs due to outdated or incorrect knowledge in …

SHIKEBLCU at SemEval-2020 Task 2: An External Knowledge-enhanced Matrix for Multilingual and Cross-Lingual Lexical Entailment

Shike Wang, Yuchen Fan, Xiangying Luo, Dong Yu

Lexical entailment recognition plays an important role in tasks like Question Answering and Machine Translation.As important branches of lexical entailment, predicting multilingual and cross-lingual lexical entailment (LE) are two subtasks of SemEval2020 Task2.In previous monolingual LE studies, researchers leverage external linguistic constraints …

cn (code pays fourni par la source)