Academic Homepage Design — Zeguan Xiao

Problem Statement

Populate the Academic Pages Jekyll template (already forked as ZeguanXiao/ZeguanXiao.github.io) with real personal and academic content so it can be published as a live academic website on GitHub Pages.

Decisions Made

Decision	Choice	Rationale
Theme	`air`	Clean, modern, white top bar with blue accents
Navigation	About (home) + Publications	No talks/teaching/portfolio content exists
Social links	Email, Google Scholar, GitHub	Only links with actual presence
CV section	Omitted	Not requested; PDF resume available if needed later

Site Configuration (`_config.yml`)

title: "Zeguan Xiao"
name: "Zeguan Xiao"
description: "PhD Candidate in Management Science and Engineering"
url: "https://ZeguanXiao.github.io"
repository: "ZeguanXiao/ZeguanXiao.github.io"
site_theme: "air"

author:
  avatar: "profile.jpg"
  name: "Zeguan Xiao"
  bio: "Final-year PhD candidate @ SUFE. Research on LLM Safety, Alignment & Robustness."
  location: "Shanghai, China"
  employer: "Shanghai University of Finance and Economics"
  email: "hainanxzg@gmail.com"
  googlescholar: "https://scholar.google.com/citations?user=rgQWhpUAAAAJ"
  github: "ZeguanXiao"

All other social fields remain blank/commented out.

Navigation (`_data/navigation.yml`)

main:
  - title: "Publications"
    url: /publications/

Only the Publications link appears in the header. The homepage (About) is always accessible at /.

Profile Photo

Copy ~/Documents/Resume-xzg/picture.jpg → images/profile.jpg, and update avatar in _config.yml to "profile.jpg".

Homepage (`_pages/about.md`)

---
permalink: /
title: "About"
author_profile: true
---

I am a **final-year** PhD candidate in Management Science and Engineering at Shanghai University of Finance and Economics (SUFE), advised by [Prof. Yun Chen](https://yunc.me/). I am also a visiting student at the Southern University of Science and Technology (SUSTech), working with [Prof. Guanhua Chen](https://ghchen.me/). My research focuses on **LLM Safety, Alignment, and Robustness**, with particular interests in LLM unlearning, uncertainty estimation, and adversarial robustness.

I am actively looking for positions (both academia and industry). Feel free to reach out at [hainanxzg@gmail.com](mailto:hainanxzg@gmail.com)!

## Research Interests

- **LLM Unlearning** — capability-preserving unlearning, robustness against relearning attacks
- **LLM Alignment** — direct alignment algorithms, reward-generation gap
- **Robustness & Uncertainty** — stress testing, uncertainty estimation via aggregated internal belief
- **LLM Safety / Jailbreak** — automatic jailbreak attack, sequential character-based attacks

## Education

- **Ph.D.** in Management Science and Engineering, Shanghai University of Finance and Economics, 2022–2026
  Advisor: [Prof. Yun Chen](https://yunc.me/)
- **M.S.** in Computer Software and Theory, Jinan University, 2018–2021
- **B.S.** in Economic Statistics, Dongbei University of Finance and Economics, 2014–2018

Publications (`_publications/`)

8 files total, one per paper, sorted by date descending. Each file uses the standard front matter with collection: publications and category: set to manuscripts (journal/conference) or conferences.

Filename	Title	Venue	Year	Category
`2026-01-01-llm-unlearning-asymmetric.md`	Modeling LLM Unlearning as an Asymmetric Two-Task Learning Problem	arXiv:2604.14808 (Under review ACL 2026)	2026	manuscripts
`2026-01-02-uncertainty-aggregated-belief.md`	Enhancing Uncertainty Estimation in LLMs with Expectation of Aggregated Internal Belief	AAAI 2026	2026	conferences
`2025-01-01-reward-generation-gap.md`	Towards Bridging the Reward-Generation Gap in Direct Alignment Algorithms	arXiv:2506.09457 (Under review ACL 2026)	2025	manuscripts
`2025-01-02-robustness-stress-testing.md`	Automatic Robustness Stress Testing of LLMs as Mathematical Problem Solvers	arXiv:2506.05038 (Under review ACL 2026)	2025	manuscripts
`2025-01-03-seqar-jailbreak.md`	SeqAR: Jailbreak LLMs with Sequential Auto-Generated Characters	NAACL 2025	2025	conferences
`2024-01-01-distract-jailbreak.md`	Distract Large Language Models for Automatic Jailbreak Attack	EMNLP 2024	2024	conferences
`2022-01-01-pruning-adaptperfusion.md`	Pruning AdaptPerFusion with Lottery Ticket Hypothesis	NAACL Findings 2022	2022	conferences
`2021-01-01-bert4gcn.md`	BERT4GCN: Using BERT Intermediate Layers to Augment GCN for Aspect-based Sentiment Classification	EMNLP 2021	2021	conferences

Each publication’s excerpt will be a one-sentence description. paperurl will link to arXiv or the ACL Anthology where available. Author lists and citation counts from Google Scholar will be included in the body of each file.

Files to Modify / Create

File	Action
`_config.yml`	Edit — personal info, theme, social links
`_data/navigation.yml`	Edit — keep only Publications
`_pages/about.md`	Edit — replace template content
`images/profile.jpg`	Add — copy from resume folder
`_publications/*.md`	Create — 8 new files, delete 5 template files
`_posts/*.md`	Delete — 5 template post files
`_talks/*.md`	Delete — 4 template talk files
`_teaching/*.md`	Delete — 2 template teaching files
`_portfolio/*.md`	Delete — 2 template portfolio files

Deployment

The repository ZeguanXiao/ZeguanXiao.github.io is already configured for GitHub Pages. After pushing, the site will be live at https://ZeguanXiao.github.io automatically via the pages-build-deployment workflow.

Zeguan Xiao