Author name cluster

Yi Zhan

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.

4 papers

1 author row

AAAI Conference 2026 Conference Paper

Forget What Has Seen: Selective Concept Unlearning in Segmentation Foundation Models

Miaozeng Du
Jiaqi Li
Sirui Pan
Yi Zhan
Guilin Qi
Yuxin Zhang
Rihui Jin
Yinjia Shu

Machine unlearning (MU) has emerged as a critical tool for removing sensitive or personal information from machine learning models, empowering individuals with the right to be forgotten. While MU has achieved success in classification and generative tasks, whether this technique can be effectively applied to segmentation foundation models remains uncertain. To address this issue, we propose an efficient method, Selective Concept Unlearning (SCU), to unlearn the segmentation capability of target concepts. SCU consists of several key aspects: (1) The Multi-level Forgetting Module, designed with a hierarchical three-level suppression strategy, including (i) distillation-level: Negative distillation steers model’s output distribution away from teacher’s correct outputs, erasing its learned concept recognition. (ii) attention-level: Attention suppression minimizes model’s attention to target regions. (iii) output-level: Directly erases predictions for the target by relabeling as background. (2) The Preservation Module ensures maintaining segmentation quality for non-target concepts. Additionally, we introduce a set of metrics to evaluate segmentation unlearning methods. Experiments demonstrate that SCU consistently outperforms existing baselines.

PDF Details DOI

EAAI Journal 2025 Journal Article

Adaptively multi-modal contrastive fusion network for molecular properties prediction

Wenyan Tang
Meng Li
Yi Zhan
Bin Chen

Molecular property prediction has become the mainstream approach for revealing the underlying mechanisms of biomedical systems with molecular representations. Existing prediction methods based on deep learning typically learn features from molecules at a specific modality or simple fusion solution, failing to consider the inconsistency, complexity, and relationships inherent in multi-modal data. To solve this issue, an adaptively multi-modal contrastive fusion network (AMCFNet) is proposed to adaptively extract the complementary features from interaction and consensus between multi-modal representations for molecular property prediction of breast cancer. The proposed model begins with a two-stream feature extractor module, which learns both one-dimensional (1D) and two-dimensional (2D) molecular representations simultaneously. The basic part of the network is the adaptively contrastive fusion module, contrastively learning features between similar and different molecules with consensus scores, which can adaptively allocate weight to fuse semantic and structural information while avoiding cognitive gaps caused by inconsistencies within multi-modal. Additionally, the final complementary molecular representation is derived by integrating 1D, 2D, and fused 1D-2D features to enhance the prediction of molecular properties in breast cancer. The proposed AMCFNet model is evaluated on five estrogen receptor alpha (ER α ) and five compound public datasets, consistently outperforming state-of-the-art baselines in classification and regression tasks of molecular property prediction including single- and multi-modal methodologies.

Details DOI

IJCAI Conference 2025 Conference Paper

CoderAgent: Simulating Student Behavior for Personalized Programming Learning with Large Language Models

Yi Zhan
Qi Liu
Weibo Gao
Zheng Zhang
Tianfu Wang
Shuanghong Shen
Junyu Lu
Zhenya Huang

Personalized programming tutoring, such as exercise recommendation, can enhance learners' efficiency, motivation, and outcomes, which is increasingly important in modern digital education. However, the lack of sufficient and high-quality programming data, combined with the mismatch between offline evaluation and real-world learning, hinders the practical deployment of such systems. To address this challenge, many approaches attempt to simulate learner practice data, yet they often overlook the fine-grained, iterative nature of programming learning, resulting in a lack of interpretability and granularity. To fill this gap, we propose a LLM-based agent, CoderAgent, to simulate students' programming processes in a fine-grained manner without relying on real data. Specifically, we equip each human learner with an intelligent agent, the core of which lies in capturing the cognitive states of the human programming practice process. Inspired by ACT-R, a cognitive architecture framework, we design the structure of CoderAgent to align with human cognitive architecture by focusing on the mastery of programming knowledge and the application of coding ability. Recognizing the inherent patterns in multi-layered cognitive reasoning, we introduce the Programming Tree of Thought (PTOT), which breaks down the process into four steps: why, how, where, and what. This approach enables a detailed analysis of iterative problem-solving strategies. Finally, experimental evaluations on real-world datasets demonstrate that CoderAgent provides interpretable insights into learning trajectories and achieves accurate simulations, paving the way for personalized programming education.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

MEgoHand: Multimodal Egocentric Hand-Object Interaction Motion Generation

Bohan Zhou
Yi Zhan
Zhongbin Zhang
Zongqing Lu

Egocentric hand-object motion generation is crucial for immersive AR/VR and robotic imitation but remains challenging due to unstable viewpoints, self-occlusions, perspective distortion, and noisy ego-motion. Existing methods rely on predefined 3D object priors, limiting generalization to novel objects, which restricts their generalizability to novel objects. Meanwhile, recent multimodal approaches suffer from ambiguous generation from abstract textual cues, intricate pipelines for modeling 3D hand-object correlation, and compounding errors in open-loop prediction. We propose MEgoHand, a multimodal framework that synthesizes physically plausible hand-object interactions from egocentric RGB, text, and initial hand pose. MEgoHand introduces a bi-level architecture: a high-level “cerebrum” leverages a vision language model (VLM) to infer motion priors from visual-textual context and a monocular depth estimator for object-agnostic spatial reasoning, while a low-level DiT-based flow-matching policy generates fine-grained trajectories with temporal orthogonal filtering to enhance stability. To address dataset inconsistency, we design a dataset curation paradigm with an Inverse MANO Retargeting Network and Virtual RGB-D Renderer, curating a unified dataset of 3. 35M RGB-D frames, 24K interactions, and 1. 2K objects. Extensive experiments across five in-domain and two cross-domain datasets demonstrate the effectiveness of MEgoHand, achieving substantial reductions in wrist translation error ( 86. 9% ) and joint rotation error ( 34. 1% ), highlighting its capacity to accurately model fine-grained hand joint structures and generalize robustly across diverse scenarios.

PDF Details