Welcome to My AI Research Blog: Multi-Agent Systems, Robotics and World Models

Welcome to my personal website and blog.

My name is Haotian Xue. I am currently a graduate student in Software Engineering at Sichuan University, where I am exploring topics at the intersection of artificial intelligence, autonomous systems, robotics, and software engineering.

I created this website as a place to organize my work, document what I learn, and share ideas that may be useful to other students, researchers, and developers.

Why I Started This Blog

Research and software development involve much more than presenting a final result. Behind every paper, experiment, or open-source project, there are usually many intermediate steps:

• understanding a new problem;
• reading papers and technical documentation;
• designing experiments;
• debugging code and environments;
• comparing alternative methods;
• recording failed attempts;
• and gradually refining an idea.

Many of these details are valuable, but they are often scattered across notebooks, repositories, documents, and temporary files. This blog gives me a more structured place to preserve them.

Writing also helps me examine whether I truly understand a topic. When an idea cannot be explained clearly, it often means that some part of the reasoning is still incomplete. For this reason, I hope this website will serve not only as a public portfolio, but also as a long-term record of my learning and research process.

What I Am Interested In

My current interests include several closely related areas.

Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning studies how multiple decision-making agents learn to cooperate, compete, communicate, and coordinate in shared environments.

I am especially interested in problems involving structured environments, decentralized decision-making, safety constraints, heterogeneous agents, and generalization to unseen scenarios.

These problems appear in many real-world systems, including intelligent transportation, robotics, autonomous vehicles, and large-scale coordination.

World Models

World models aim to learn compact representations of an environment and predict how that environment may evolve.

A useful world model can help an intelligent agent reason about future states before taking an action. This creates opportunities for planning, imagination-based learning, sample-efficient reinforcement learning, and long-horizon decision-making.

I am interested in how world models can be extended from single-agent settings to multi-agent and embodied systems.

Robotics and Autonomous Systems

Robotics connects perception, planning, control, and interaction with the physical world.

I am particularly interested in autonomous systems that must operate under uncertainty, respond to changing environments, and coordinate with other agents. This includes aerial robots, intelligent vehicles, and other embodied platforms.

Intelligent Transportation

Transportation systems naturally involve many interacting agents, shared resources, safety requirements, and complex operational constraints.

I am interested in how learning-based methods can be combined with graph representations, domain knowledge, planning algorithms, and safety rules to improve coordination and decision-making in these systems.

Open-Source Software

Open-source development is also an important part of my learning process.

Implementing ideas, organizing code, reproducing experiments, and documenting software are essential skills for turning research concepts into usable systems. I plan to share development records, environment setup notes, debugging experiences, and lessons learned from my projects.

What I Will Share Here

The content of this blog will mainly include:

• research notes on artificial intelligence and reinforcement learning;
• paper reading notes and method explanations;
• technical tutorials and environment setup guides;
• experiment design and implementation records;
• project development summaries;
• debugging notes and solutions to practical problems;
• reflections on research, learning, and open-source development.

Some posts will focus on theoretical ideas, while others will be more practical and code-oriented. I hope to keep the explanations clear and reproducible whenever possible.

My Background

Before beginning my graduate study at Sichuan University, I completed a bachelor’s degree in Computer Science and Technology at the Civil Aviation University of China.

My previous studies and projects involved software engineering, systems programming, computer networks, and open-source development. These experiences gradually led me toward artificial intelligence, autonomous systems, and multi-agent decision-making.

I am still learning, and this website will evolve together with my research interests and technical experience.

Looking Ahead

This is the first post on the website, but it is not intended to remain only a simple welcome page.

Over time, I hope the blog will become a searchable collection of useful research notes, technical explanations, project records, and lessons from both successful and unsuccessful experiments.

You can also find my open-source work on GitHub and some of my earlier technical writing on my CSDN Blog.

Thank you for visiting.