Author name cluster

Yiwei Chen

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

6 papers

2 author rows

AAAI Conference 2026 Conference Paper

Flow-Based Knowledge Transfer for Efficient Large Model Distillation

Xinye Yang
Junhao Wang
RuiLi
Haosen Sun
Xuesheng Zhang
Zebang Liu
Gaochao Xu
Yiwei Chen

Traditional knowledge distillation relies on simple MSE or KL divergence losses that fail to capture the complex distributional relationships between teacher and student model representations. We propose FlowDistill, a novel distillation framework that employs normalizing flows to model and transfer the intricate knowledge distributions from teacher to student models. Our approach introduces three key innovations: (1) Invertible Knowledge Mapping using continuous normalizing flows (CNFs) to learn bijective transformations between teacher and student representation spaces, enabling precise knowledge transfer without information loss, (2) Flow-Guided Progressive Distillation that gradually increases the complexity of knowledge transfer by learning hierarchical flow transformations from simple to complex distributions, and (3) Conditional Flow Networks that adapt knowledge transfer based on input context and task requirements. Unlike previous diffusion-based distillation methods such as DiffKD that suffer from computational overhead due to iterative denoising processes and information loss during noise addition, our flow-based approach provides exact invertible transformations with significantly reduced computational cost. Extensive experiments on ImageNet classification, COCO object detection, and Cityscapes semantic segmentation demonstrate that FlowDistill achieves superior performance with 2.1% accuracy improvement over DiffKD on ResNet-34 to ResNet-18 distillation while reducing inference time by 3.5×. Our method establishes new state-of-the-art results across multiple distillation benchmarks and provides theoretical guarantees for lossless knowledge transfer through invertible flow transformations.

PDF Details DOI

EAAI Journal 2026 Journal Article

Quantum-inspired neural networks with stochastic dynamics for multimodal sentiment analysis and sarcasm detection

Kehuan Yan
Peichao Lai
Xianghan Zheng
Yang Yang
Yi Ren
Tuyatsetseg Badarch
Yiwei Chen

Quantum-inspired neural networks have demonstrated strong potential in modeling non-classical phenomena in cognitive tasks, particularly in multimodal sentiment analysis, marking a significant advancement over traditional models. However, existing multimodal quantum-inspired neural networks fall short in fully modeling the multimodal density matrix, typically relying on simplistic neural mappings to represent quantum entanglement. This lack of explicit physical constraints, particularly those governing open quantum system dynamics, limits both the interpretability and performance. To address this limitation, we propose a novel framework grounded in quantum stochastic dynamics, introducing two quantum-inspired neural networks, which model the evolution of multimodal data as Markovian and non-Markovian open quantum systems, respectively. This approach enables the simulation of quantum system evolution to capture rich non-classical interactions between modalities. The resulting entangled multimodal density matrix is then measured through quantum projections to extract high-level features for downstream sentiment analysis and sarcasm detection. Extensive experiments on benchmark bimodal and trimodal datasets demonstrate that our models consistently outperform state-of-the-art traditional baselines, large-scale language models and quantum-inspired neural networks. Ablation studies confirm the critical role of quantum stochastic dynamics in performance gains. Furthermore, we enhance the interpretability by tracking the evolution of the density matrix using von-Neumann entanglement entropy as a quantitative metric, providing deeper insight into the internal mechanisms of the model.

Details DOI

NeurIPS Conference 2025 Conference Paper

The Fragile Truth of Saliency: Improving LLM Input Attribution via Attention Bias Optimization

Yihua Zhang
Changsheng Wang
Yiwei Chen
Chongyu Fan
Jinghan Jia
Sijia Liu

Input saliency aims to quantify the influence of input tokens on the output of large language models (LLMs), which has been widely used for prompt engineering, model interpretability, and behavior attribution. Despite the proliferation of saliency techniques, the field lacks a standardized and rigorous evaluation protocol. In this work, we introduce a stress-testing framework inspired by the needle-in-a-haystack (NIAH) setting to systematically assess the reliability of seven popular input saliency methods. Our evaluation reveals a surprising and critical flaw: existing methods consistently assign non-trivial importance to irrelevant context, and this attribution error worsens as input length increases. To address this issue, we propose a novel saliency method based on Attention Bias Optimization (ours), which explicitly optimizes the attention bias associated with each input token to quantify its causal impact on target token generation. ABO robustly outperforms existing methods by 10\sim30% in saliency accuracy across diverse NIAH tasks, maintains effectiveness up to 10K-token prompts, and enables practical applications including zero-shot detoxification, sentiment steering, and reasoning-error correction. Our findings highlight the limitations of prevalent attribution methods and establish ABO as a principled alternative for accurate token attribution.

PDF Details

NeurIPS Conference 2023 Conference Paper

BadTrack: A Poison-Only Backdoor Attack on Visual Object Tracking

Bin Huang
Jiaqian Yu
Yiwei Chen
Siyang Pan
Qiang Wang
Zhi Wang

Visual object tracking (VOT) is one of the most fundamental tasks in computer vision community. State-of-the-art VOT trackers extract positive and negative examples that are used to guide the tracker to distinguish the object from the background. In this paper, we show that this characteristic can be exploited to introduce new threats and hence propose a simple yet effective poison-only backdoor attack. To be specific, we poison a small part of the training data by attaching a predefined trigger pattern to the background region of each video frame, so that the trigger appears almost exclusively in the extracted negative examples. To the best of our knowledge, this is the first work that reveals the threat of poison-only backdoor attack on VOT trackers. We experimentally show that our backdoor attack can significantly degrade the performance of both two-stream Siamese and one-stream Transformer trackers on the poisoned data while gaining comparable performance with the benign trackers on the clean data.

PDF Details

ICML Conference 2021 Conference Paper

Learning Generalized Intersection Over Union for Dense Pixelwise Prediction

Jiaqian Yu
Jingtao Xu
Yiwei Chen
Weiming Li
Qiang Wang 0023
ByungIn Yoo
Jae-Joon Han

Intersection over union (IoU) score, also named Jaccard Index, is one of the most fundamental evaluation methods in machine learning. The original IoU computation cannot provide non-zero gradients and thus cannot be directly optimized by nowadays deep learning methods. Several recent works generalized IoU for bounding box regression, but they are not straightforward to adapt for pixelwise prediction. In particular, the original IoU fails to provide effective gradients for the non-overlapping and location-deviation cases, which results in performance plateau. In this paper, we propose PixIoU, a generalized IoU for pixelwise prediction that is sensitive to the distance for non-overlapping cases and the locations in prediction. We provide proofs that PixIoU holds many nice properties as the original IoU. To optimize the PixIoU, we also propose a loss function that is proved to be submodular, hence we can apply the Lovász functions, the efficient surrogates for submodular functions for learning this loss. Experimental results show consistent performance improvements by learning PixIoU over the original IoU for several different pixelwise prediction tasks on Pascal VOC, VOT-2020 and Cityscapes.

Details

AAAI Conference 2021 System Paper

OPRA: An Open-Source Online Preference Reporting and Aggregation System

Yiwei Chen
Jingwen Qian
Junming Wang
Lirong Xia
Gavriel Zahavi

We introduce the Online Preference Reporting and Aggregation (OPRA) system, an open-source online system that aims at providing support for group decision-making. We illustrate OPRA’s distinctive features: UI for reporting rankings with ties, comprehensive analytics of preferences, and group decision-making in combinatorial domains. We also discuss our work in an automatic mentor matching system. We hope that the open-source nature of OPRA will foster development of computerized group decision support systems.

PDF Details