Zixuan Jiang

Zixuan Jiang

Xi'an Jiaotong University

AndrewBot Knowledge Base (Refined)

Version: 2026-05-09 Language coverage: English + Chinese Primary owner: Zixuan Jiang (Andrew)

1) Profile

Basic identity

  • Name: Zixuan Jiang (Andrew) / 蒋梓轩
  • Affiliation: College of Artificial Intelligence, Xi’an Jiaotong University
  • Location: Xi’an, China
  • Email: andrewjiang@stu.xjtu.edu.cn

Research focus

  • EN: Multimodal intelligence across vision, speech, and natural language; interests in omni-modal agents and human-computer interaction.
  • ZH: 研究方向包括视觉、语音、自然语言的多模态智能，并关注全模态智能体和人机交互。
  • Google Scholar: https://scholar.google.com/citations?user=hu8iqsMAAAAJ
  • GitHub: https://github.com/AnXMuy
  • ResearchGate: https://www.researchgate.net/profile/Zixuan-Jiang-8?ev=hdr_xprf
  • ORCID: https://orcid.org/my-orcid?orcid=0009-0008-2706-553X

3) Education & Internship

Education

  • 2023.09 - now: Artificial Intelligence Experimental Class, Xi’an Jiaotong University
  • 2021.09 - now: Young Gifted Program, Xi’an Jiaotong University

Internship / research groups

  • Xi’an Jiaotong University, Associate Professor Xiangyong Cao’s research group
  • Shanghai Jiao Tong University, X-LANCE Lab

4) Recent News

  • 2025.12: Pre-admission offer from Shanghai Innovation Institute (2026 Golden Autumn Camp)
  • 2025.10: National Scholarship confirmed
  • 2025.10: DescribeEarth open-sourced
  • 2025.09: DescribeEarth released on arXiv

5) Publications (Core Facts)

Remote sensing

  1. DescribeEarth: Describe Anything for Remote Sensing Images (arXiv 2025)
    • Link: https://arxiv.org/abs/2509.25654
    • Positioning: describe-anything for remote sensing images
    • Resources: code + dataset + benchmark released
  2. Annotation-Free Open-Vocabulary Segmentation for Remote-Sensing Images (arXiv 2025)
    • Link: https://arxiv.org/abs/2508.18067
    • Positioning: annotation-free OV segmentation in remote sensing
  3. Advances in Open Vocabulary Perception for Remote Sensing Images (Journal of Image and Graphics 2026)
    • Link: https://www.cjig.cn/en/article/doi/10.11834/jig.260163/
    • Positioning: open-vocabulary perception review/progress in remote sensing

Audio interaction

  1. Interactive ASR: Towards Human-Like Interaction and Semantic Coherence Evaluation for Agentic Speech Recognition (arXiv 2026)
    • Link: https://arxiv.org/abs/2604.09121
    • Positioning: interactive ASR; LLM-as-judge based semantic success evaluation

6) Honors and Competitions

Honors

  • 2025.10: National Scholarship
  • 2025.10: Outstanding Student, Xi’an Jiaotong University
  • 2024.11: Shuiyou First-Class Scholarship (Top 4 in university)
  • 2024.11: Outstanding Student, Xi’an Jiaotong University
  • 2023.11: Third-Class Scholarship, Xi’an Jiaotong University

Competitions

  • 2025: Second Prize, National Laser Radar Conference Point Cloud Intelligent Analysis Competition
  • 2025: Silver Medal, ICPC Shaanxi Provincial Programming Contest
  • 2025: Excellence Award (Finalist), AI+ College Innovation Program
  • 2024: Bronze Medal, ICPC Shaanxi Provincial Programming Contest
  • 2023-2025: 3x First Prize, National College Student Mathematical Modeling Competition (Shaanxi Division)
  • 2024: Meritorious Winner, MCM (USA)
  • 2023: Honorable Prize, MCM (USA)

7) FAQ-style Snippets (for better retrieval)

Q: Andrew 的主要科研方向是什么？

A: 多模态智能（视觉/语音/自然语言），并关注全模态智能体与人机交互。

Q: What are Andrew’s representative papers?

A: DescribeEarth (arXiv 2025), Annotation-Free OV Segmentation for RS (arXiv 2025), Interactive ASR (arXiv 2026), and an open-vocabulary remote sensing review (JIG 2026).

Q: Andrew 在哪里学习和实习？

A: 西安交通大学人工智能学院；曾在西交曹翔勇老师课题组与上交 X-LANCE 实验室开展研究。

Q: How can I follow Andrew’s academic updates?

A: Check Google Scholar, GitHub, and ResearchGate links in this knowledge base.

8) Source Inventory (bot_memory)

The following files are expected under ../bot_memory and used by extraction pipeline:

  • 2508.18067v1.pdf
  • 2509.25654v1.pdf
  • 2604.09121v3.pdf
  • 2604.11998v1.pdf
  • Advances in open vocabulary perception for remote sensing images_NormalPdf.pdf
  • CV.pdf

9) Bot Answering Policy

  • Prioritize explicit facts from this file and extracted PDF knowledge.
  • If evidence is insufficient, clearly say uncertainty instead of guessing.
  • Prefer concise, factual, and source-aware answers.