---
title: "S-DAG: A Subject-Based Directed Acyclic Graph for Multi-Agent Heterogeneous Reasoning"
collection: publications
category: conferences
permalink: /publication/2026-s-dag
excerpt: "We propose S-DAG, a subject-based directed acyclic graph framework for multi-agent heterogeneous reasoning that enables structured collaboration among diverse LLM agents."
date: 2026-01-20
venue: "AAAI Conference on Artificial Intelligence"
paperurl: "https://arxiv.org/abs/2511.06727"
citation: 'Dong, Jiangwen, Zehui Lin, Wanyu Lin, and Mingjin Zhang. "S-DAG: A Subject-Based Directed Acyclic Graph for Multi-Agent Heterogeneous Reasoning." AAAI 2026.'
---

We propose S-DAG, a subject-based directed acyclic graph framework for multi-agent heterogeneous reasoning that enables structured collaboration among diverse LLM agents.

---
title: "HybridFlow: Adaptive Task Scheduling for Fast and Token-Efficient LLM Inference in Edge-Cloud Collaboration"
collection: publications
category: conferences
permalink: /publication/2026-hybridflow
excerpt: "HybridFlow introduces an adaptive task scheduling framework for efficient LLM inference across edge-cloud environments, optimizing both latency and token usage."
date: 2026-01-20
venue: "IEEE International Conference on Distributed Computing Systems"
paperurl: "https://arxiv.org/abs/2512.22137"
citation: 'Dong, Jiangwen, Jiayu Li, and Wanyu Lin. "HybridFlow: Adaptive Task Scheduling for Fast and Token-Efficient LLM Inference in Edge-Cloud Collaboration." ICDCS 2026.'
---

HybridFlow introduces an adaptive task scheduling framework for efficient LLM inference across edge-cloud environments, optimizing both latency and token usage.

---
permalink: /
title: ""
author_profile: false
redirect_from:
  - /about/
  - /about.html
---

<!-- Hero Section -->
<div class="hero-section">
  <div class="hero-content" style="position: relative; z-index: 1; max-width: 1200px; margin: 0 auto; padding: 0 2rem; width: 100%;">
    <div style="max-width: 60%;">
      <h1>Jiangwen Dong</h1>
      <p class="subtitle">PhD Student · PolyU · Graph-Driven Multi-Agent Systems</p>
      <p class="mission">I develop graph-driven multi-agent systems that enable efficient, adaptive, and distributed LLM reasoning across edge–cloud environments.</p>
      <div class="hero-social-links">
        <a href="https://scholar.google.com/citations?user=Kl-dTYAAAAAJ&hl=en&oi=ao" target="_blank" aria-label="Google Scholar"><i class="fas fa-graduation-cap"></i></a>
        <a href="https://github.com/Jiangwen-Dong" target="_blank" aria-label="GitHub"><i class="fab fa-github"></i></a>
        <a href="mailto:jiangwen.dong@connect.polyu.hk" aria-label="Email"><i class="fas fa-envelope"></i></a>
      </div>
    </div>
  </div>
  <canvas id="hero-graph-canvas"></canvas>
</div>

<!-- Research Pillars Section -->
<div class="section">
  <div style="max-width: 1200px; margin: 0 auto; padding: 0 2rem;">
    <h2 class="section-title">Research</h2>
    <div class="research-grid">
      <div class="research-card">
        <div class="research-card-icon">🕸️</div>
        <h3>Structured</h3>
        <p>Graph-based reasoning methods for organizing and guiding LLM inference.</p>
      </div>
      <div class="research-card">
        <div class="research-card-icon">🤖</div>
        <h3>Collaborative</h3>
        <p>Multi-agent LLM systems that coordinate, communicate, and share knowledge.</p>
      </div>
      <div class="research-card">
        <div class="research-card-icon">⚡</div>
        <h3>Efficient</h3>
        <p>Edge–cloud computation paradigms for distributed and resource-aware AI.</p>
      </div>
      <div class="research-card">
        <div class="research-card-icon">🔄</div>
        <h3>Adaptive</h3>
        <p>Continual learning and personalization in dynamic multi-agent settings.</p>
      </div>
    </div>
  </div>
</div>

<!-- Recent Publications Section -->
<div class="section section-alt">
  <div style="max-width: 1200px; margin: 0 auto; padding: 0 2rem;">
    <h2 class="section-title">Selected Publications</h2>

    <div class="pub-card">
      <div class="pub-venue-badge">AAAI 2026</div>
      <div class="pub-content">
        <h3><a href="/publication/2026-s-dag">S-DAG: A Subject-Based Directed Acyclic Graph for Multi-Agent Heterogeneous Reasoning</a></h3>
        <p class="pub-authors"><span class="highlight">Jiangwen Dong</span>, Zehui Lin, Wanyu Lin, Mingjin Zhang</p>
        <p class="pub-abstract">A subject-based DAG framework enabling structured collaboration among diverse LLM agents for complex reasoning tasks.</p>
        <div class="pub-links">
          <a href="https://arxiv.org/abs/2511.06727" target="_blank">Paper →</a>
        </div>
      </div>
    </div>

    <div class="pub-card">
      <div class="pub-venue-badge">ICDCS 2026</div>
      <div class="pub-content">
        <h3><a href="/publication/2026-hybridflow">HybridFlow: Adaptive Task Scheduling for Fast and Token-Efficient LLM Inference in Edge-Cloud Collaboration</a></h3>
        <p class="pub-authors"><span class="highlight">Jiangwen Dong</span>, Jiayu Li, Wanyu Lin</p>
        <p class="pub-abstract">An adaptive task scheduling framework for efficient LLM inference across edge-cloud environments.</p>
        <div class="pub-links">
          <a href="https://arxiv.org/abs/2512.22137" target="_blank">Paper →</a>
        </div>
      </div>
    </div>

    <p style="margin-top: 2rem;"><a href="/publications/" style="font-weight: 600;">View all publications →</a></p>
  </div>
</div>

---
permalink: /research/
title: "Research"
author_profile: true
---

I develop graph-driven multi-agent systems that enable efficient, adaptive, and distributed LLM reasoning across edge–cloud environments. My research integrates graph intelligence, multi-agent collaboration, and efficient distributed systems.

## Research Areas

<div class="research-grid" style="margin-top: 2rem;">
  <div class="research-card">
    <div class="research-card-icon">🕸️</div>
    <h3>Structured Reasoning</h3>
    <p>Graph-based reasoning methods organize LLM inference into structured, interpretable workflows. By representing reasoning steps as nodes in a directed graph, we can guide multi-agent systems through complex problem-solving with explicit dependencies and subject-based decomposition.</p>
    <p style="margin-top: 1rem; font-size: 0.85rem;"><strong>Related:</strong> <a href="/publication/2026-s-dag">S-DAG (AAAI 2026)</a></p>
  </div>

  <div class="research-card">
    <div class="research-card-icon">🤖</div>
    <h3>Collaborative Multi-Agent Systems</h3>
    <p>Multi-agent LLM systems coordinate specialized agents to solve problems beyond the capability of any single model. My work focuses on heterogeneous agent collaboration, where agents with different capabilities and roles communicate through structured protocols.</p>
    <p style="margin-top: 1rem; font-size: 0.85rem;"><strong>Related:</strong> <a href="/publication/2026-s-dag">S-DAG (AAAI 2026)</a></p>
  </div>

  <div class="research-card">
    <div class="research-card-icon">⚡</div>
    <h3>Efficient Distributed Inference</h3>
    <p>Edge–cloud collaboration enables efficient LLM inference by distributing computation across heterogeneous devices. I design adaptive scheduling and token-efficient strategies that minimize latency while respecting resource constraints at the edge.</p>
    <p style="margin-top: 1rem; font-size: 0.85rem;"><strong>Related:</strong> <a href="/publication/2026-hybridflow">HybridFlow (ICDCS 2026)</a></p>
  </div>

  <div class="research-card">
    <div class="research-card-icon">🔄</div>
    <h3>Adaptive Learning</h3>
    <p>Continual and personalized learning in multi-agent settings allows systems to adapt to new tasks, user preferences, and changing environments without retraining from scratch. This includes federated personalization and online adaptation strategies.</p>
  </div>
</div>

## Active Projects

- **S-DAG Framework** — Subject-based DAG for multi-agent reasoning. <span style="color: var(--color-teal); font-weight: 600;">Accepted at AAAI 2026</span>
- **HybridFlow** — Adaptive edge-cloud LLM inference scheduling. <span style="color: var(--color-teal); font-weight: 600;">Accepted at ICDCS 2026</span>

## Collaborations

I am fortunate to work with my advisor <a href="https://www.comp.polyu.edu.hk/~wanyulin/" target="_blank">Prof. Wanyu Lin</a> and collaborators at The Hong Kong Polytechnic University.

<div class="cv-layout">
  <div class="cv-sidebar">
    <!-- sidebar content -->
  </div>
  <div class="cv-main">
    <!-- main content -->
  </div>
</div>