Frontier of Artificial Network
A Series of Invited Talks @ FAN Group, CityU

The Upcoming Talk
基于冻结特征空间决策面搜索的领域自适应方法

哈尔滨工业大学

Date: July 28, 2026 (Tue)
Time: 14:30 (HKT)
Zoom: 801 137 0362

Biography: 现任哈尔滨工业大学数学学院副研究员。本科和博士毕业于北京大学地球与空间科学学院,导师为宁杰远教授。研究领域为人工智能和地球物理。在TPAMI、Journal of Geophysical Research: Solid Earth、Journal of Geophysical Research: Machine Learning and Computation、Science China Earth Sciences、科学通报、地球物理学报等期刊上发表论文17篇,谷歌学术引用275次。主持国自然青C、中国博士后面上、深地国家重点专项专题等项目。


Abstract: 域偏移是人工智能模型从实验室走向真实场景的障碍。训练数据与实际数据常因设备、环境及采样条件不同而分布不一致,导致模型跨域性能下降。传统无监督领域自适应常通过调整深层特征缩小域差异,但需更新大量参数,计算开销高,且难以解释和控制。我们发现当预训练模型已形成稳定特征空间时,可将领域自适应转化为固定空间中的决策边界优化,并提出了Feature-Space Planes Searcher领域自适应框架:无需重塑特征,只需利用目标域已有的结构先验与几何模式,如类内紧凑、类间分离、边界避开高密度区域,搜索更符合目标分布的决策平面。实验表明,仅调整决策边界也能实现高效、稳定、可解释的跨域迁移。该框架适用于自然图像、蛋白质结构预测、遥感分类和地震检测等多种场景,并为大模型低成本可靠部署提供新路径。

Previous Talks

Date Speaker Title Materials
July 20, 2026 Adeel Razi Synthetic Biological Intelligence: Learning, Adaptation, and Brain-Inspired AI [Poster]
July 14, 2026 Ran Cheng From AI Coding to Software Evolution: Towards Trustworthy Evolutionary Agent Systems [Poster]
July 13, 2026 Yongqiang Cai Neural Networks as a Discretization of Dynamical Systems [Poster]
July 11, 2026 Renfeng Peng Geometric Methods for Low-Rank Tensor Optimization Problems [Poster]
June 25, 2026 Hongkang Li Towards a Mathematical Theory of Foundation Models [Poster]
June 18, 2026 Shuai Zhang Towards Theoretical Foundations of Foundation Models: Feature Learning, Optimization Dynamics, and Generalization [Poster]
June 11, 2026 Miao Zhang Large Model Compression and Inference Acceleration [Poster]
June 8, 2026 Jinwei Zhang Lesion-Centric Quantitative Neuroimaging: Integrating MRI Innovation and AI-Driven Analysis for Neurodegeneration [Poster]
June 5, 2026 Xiaolin Hu Brain-Inspired Speech Separation Models [Poster]
June 5, 2026 Lisha Chen Principled Preference-Guided Multi-Objective Learning: Navigating and Uniformly Traversing the Pareto Front [Poster]
May 28, 2026 Shirui Pan Boosting Large Language Model Reasoning with Knowledge Graphs [Poster]
May 13, 2026 Xiaojuan Qi 3D Representations for World Modeling and Generation [Poster]
May 7, 2026 Yuantian Miao The Audio Auditor: User-level Membership Inference in Internet of Things Voice Services [Poster]
Apr 29, 2026 Han Zhong Toward Principled Reinforcement Learning: From Statistical Complexity to Representation Complexity [Poster]
Apr 16, 2026 Yudong Zhang AI-Integrated Colorectal Cancer Research: Challenges, Progress and Innovation [Poster]
Apr 9, 2026 Bin Gao Low-rank Optimization Through the Lens of Geometry [Poster]
Mar 31, 2026 Zhuo Sun Multilevel Control Functional [Poster]
Mar 25, 2026 Jiancheng Yang Scaling Medical AI Without Scaling Cost in the Era of Generative AI [Poster]
Mar 24, 2026 Yang Cao Differential Privacy in LLM Fine-Tuning: What It Protects, What It Costs, and What It Doesn’t [Poster]
Mar 18, 2026 Zijun Cui AI + Knowledge: Unleashing the Power of Domain Knowledge for Advanced Artificial Intelligence [Poster]
Mar 12, 2026 Chenxi Yuan Enhance Prediction of Alzheimer’s Disease with Generative AI [Poster]
Mar 3, 2026 Ren Wang Robustness Through Collective Intelligence [Poster]
Feb 28, 2026 Huiyu Zhou Constructing Masterpieces From Missing Pieces [Poster]
Feb 10, 2026 Guibo Luo Benchmarking Multi-Party Privacy Computing and Exploring New Collaboration Paradigms [Poster]
Feb 5, 2026 Zeyu Zhang Bridging Scene Understanding and Motion Generation in Robot Manipulation [Poster]
Jan 28, 2026 Md Sajid Interpretable and Robust Randomized Neural Networks for Real-World Learning [Poster]
Jan 21, 2026 Bing Yang Shape-Aware Deep Learning for AS-OCT Analysis: Segmentation and Structural Uncertainty [Poster]
Jan 14, 2026 Mengdi Zhao Simulating Biological Intelligence: Bridging High-Fidelity Neuronal Modeling with Embodied Agents [Poster]
Jan 7, 2026 Xiangyu Chang Research on Efficient and Fair Data Element Pricing Mechanisms [Poster]
Dec 19, 2025 Shujian Huang Cross-lingual Knowledge Learning and Reasoning in Large Language Models [Poster]
Dec 17, 2025 Haotong Qin Extreme Discretization: Towards Efficient Intelligence and Systems in the Scaling Era [Poster]
Dec 3, 2025 Ningning Ding From Fair Unlearning Algorithms to Incentive-Compatible Mechanisms in Federated Unlearhing [Poster]
Oct 23, 2025 Raian Ali AI Design vs. Human Attitude, Learning, and Dependency [Poster]
Oct 22, 2025 Yuwen Li Higher Order Approximation Error Bounds for ReLU Neural Networks in Korobov Space [Poster]
Oct 13, 2025 Weijie Su The ICML 2023 Ranking Experiment: Empirical Performance and Analysis of the Isotonic Mechanism [Poster]
Sep 17, 2025 Fanghui Liu Bridging Theory and Practice: One-step Full Gradient Can Suffice for Low-rank Fine-tuning in LLMs [Poster]
Aug 20, 2025 Fan Yang RNA Recognition and Targeted Degradation: Mechanisms and Engineering Strategies based on RNA-Binding Domains (RBDs) [Poster]
Aug 4, 2025 Tao Luo The Theory of Parameter Condensation in Neural Networks [Poster]
Jul 29, 2025 Hangxin Liu Embodied Mobile Manipulation: Trajectory Optimization vs. Diffusion [Poster]
Jul 18, 2025 Yide Liu Constructing High-performance Robotic Insects With Origami Transmission Mechanism [Poster]
Jul 15, 2025 Hai Dong Mobile Edge Intelligence: When AI Meets Mobile Edge Computing [Poster]
Jun 30, 2025 Guangyi Chen Causal Representation Learning for Visual Understanding [Poster]

Organizers

Fenglei Fan, Assistant Professor in the Department of Data Science at the City University of Hong Kong
Shuren Qi, Postdoctoral Fellow in the Department of Data Science at the City University of Hong Kong