Haichao Zhang

About Me关于我

I am a Ph.D. student in the School of AI and Advanced Computing (AIAC) atXi'an Jiaotong-Liverpool University (XJTLU), supervised byProf. Jia Wang. I am pursuing a Ph.D. in Computer Science and Software Engineering, with the degree awarded by the University of Liverpool.我是西交利物浦大学（XJTLU）人工智能与先进计算学院（AIAC）的博士研究生，导师为王佳教授。 我攻读计算机科学与软件工程博士学位，学位由利物浦大学授予。

I study trustworthy and controllable recommender systems at the intersection of large language models, retrieval-augmented generation, machine unlearning, and model editing.我的研究聚焦可信与可控推荐系统，探索大语言模型、检索增强生成、机器遗忘和模型编辑的交叉问题。

01

Recommendation Unlearning推荐遗忘

02

LLMs for Recommendation大模型推荐

03

Responsible Personalization负责任的个性化

News近期动态

Experience & Education工作与教育经历

Education教育经历

XJTLUUniversity of Liverpool

Xi'an Jiaotong-Liverpool University西交利物浦大学

Ph.D. in Computer Science and Software Engineering计算机科学与软件工程博士

School of AI and Advanced Computing · Degree awarded by the University of Liverpool人工智能与先进计算学院 · 学位由利物浦大学授予

Industry Experience工作经历

Alibaba

Alibaba · Beijing阿里巴巴 · 北京

Development Engineer开发工程师

Built audience and recommendation data infrastructure for the DMP platform, including engine control, observability, and large-scale inspection and attribution systems.参与 DMP 人群与推荐数据平台建设，负责计算引擎管控、可观测性及大规模巡检归因系统。

Dingfu Data

Dingfu Data Technology · Beijing鼎富智能科技 · 北京

Development Engineer开发工程师

Worked on computer-vision algorithms and image-processing pipelines, with a focus on watermark removal.从事计算机视觉算法与图像处理流程研发，主要研究图像水印去除。

Publications学术论文

Google Scholar profileGoogle Scholar 主页 ↗
IEEE ICDM 2025

Customized Retrieval-Augmented Generation with LLM for Debiasing Recommendation Unlearning

Haichao Zhang, Chong Zhang, Peiyu Hu, Shi Qiu, Jia Wang

2025 IEEE International Conference on Data Mining (ICDM), 1702-1711, 2025

A retrieval-augmented generation framework for user-level recommendation unlearning that limits collateral effects on non-target users while preserving recommendation quality.面向用户级推荐遗忘的检索增强生成框架，在保留推荐质量的同时，减少遗忘操作对非目标用户的连带影响。

[Paper论文][Preprint预印本][Code代码][Project项目]Cited by 2引用 2

Ongoing Research在研工作

Submitted and early-stage work is intentionally introduced only at a high level.为避免泄露尚未发表的研究思路，此处仅提供概括性介绍。

Under review在投 / 在研

Teaching to Forget: Dual-Teacher Distilled Prompt-Tuning for Efficient Recommendation Unlearning

A lightweight study of on-demand recommendation unlearning that aims to preserve utility for unaffected users.一项轻量级按需推荐遗忘研究，目标是在移除指定影响的同时保留未受影响用户的推荐效用。

Under review在投 / 在研

Controllable Generative Recommendation via Guided Token Refinement

A controllable generative recommendation study focused on more reliable alignment between user intent and generated recommendations.一项可控生成式推荐研究，关注用户意图与生成推荐结果之间更可靠的对齐。

Under review在投 / 在研

Personalized Conformity Disentanglement for Debiased Recommendations

A debiasing study that distinguishes personal preference signals from conformity effects for more faithful personalization.一项推荐去偏研究，通过区分个体偏好与从众效应，实现更忠实的个性化。

Under review在投 / 在研

Explain-then-Forget: Causal Explanation-based Unlearning for Efficient and Precise Recommendation

An efficient recommendation unlearning study investigating how causal explanations can guide precise forgetting.一项高效推荐遗忘研究，探索因果解释如何指导更加精确的遗忘。

Under review在投 / 在研

Dual-Rate User Semantic Memory for LLM-Enhanced Sequential Recommendation

A dual-rate semantic memory for efficient LLM-enhanced sequential recommendation, designed to retain rich user semantics with lightweight online serving.一种面向高效大模型增强序列推荐的双速率语义记忆，在轻量在线服务中保留丰富用户语义。

Open Source开源项目

axiom-quant

A local-first open-source A-share quantitative research and batch backtesting platform.本地优先的开源 A 股量化研究与批量回测平台。

Python ↗

senpai-skill

An open-source AI research mentor skill for shaping topics, planning experiments, and improving academic work.开源 AI 科研导师 Skill，支持选题打磨、实验规划与学术工作改进。

Python ↗

PaperReader

A compact home for structured paper reading, research notes, and reusable academic insights.用于结构化论文阅读、研究笔记与学术知识沉淀的轻量开源项目。

TypeScript ↗

Selected Awards部分荣誉

  • First Prize, National Railway Transportation Student Thesis Competition全国轨道交通高校学生优秀论文一等奖
  • Second Prize, Jiangxi Computer Works and Internet Innovation Competition江西省计算机作品赛暨“互联网+”创新创业大赛二等奖
  • Third Prize, National Internet Transportation Innovation Competition全国高校“互联网交通”创新创业大赛三等奖
  • First Prize, ECJTU ACM Programming Competition华东交通大学 ACM 程序设计竞赛一等奖