Xue Yao / 姚雪Curriculum vitae ↗

ACTIVE INFERENCE × INTELLIGENT TRANSPORTATION

Understanding behavior.
Rethinking
traffic intelligence.

I’m Xue Yao. I explore active inference as a framework for learning, decision-making, and coordination in transportation systems.

My work connects human driving heterogeneity with collective traffic dynamics and interpretable AI.

POSTDOCTORAL RESEARCHERTU Delft
VISITING SCHOLARUC Berkeley
RESEARCH IN FOCUS01 / ACTIVE INFERENCE
A perception–action loop for intelligent traffic systemsAn illustrative diagram connects observations, beliefs and actions around a transportation network.OBSERVATIONSBELIEFSACTIONSFrom individual decisions to collective dynamics.
PERCEIVE → LEARN → ACTConceptual illustration
Active inference · Human driving behavior · Traffic flow theory · Mixed autonomy

01 / RESEARCH

From human decisions
to intelligent systems.

How can transportation systems learn, adapt, and coordinate—while accounting for the people within them?

01
PRIMARY DIRECTION

Active inference &
intelligent traffic management

I investigate active inference as a unified framework for behavior modeling, learning, and decentralized decision-making and control—from individual road users to traffic networks.

My current postdoctoral research contributes to AiMTT’s work on AI-enabled multimodal traffic management at TU Delft.

Related work ↗
Perception↓Learning↓Decision↓Coordination
02

Driving behavior
& heterogeneity

Understanding inter- and intra-driver variability through interpretable, data-driven models of human behavior, decision-making, and adaptation.

Related work ↗
03

Traffic dynamics
& mixed autonomy

Investigating how interactions between human-driven and automated vehicles shape stability, congestion, safety, and collective traffic performance.

Related work ↗

03 / PUBLICATIONS

The research record.

Google Scholar ↗

04 / TEACHING & COMMUNITY

Sharing knowledge.
Building connections.

Teaching & supervision

Delft University of Technology

  • AI in Traffic and Transport EngineeringLecturer · CTB3395 · 2026/27 (Q1)
  • Methods in Traffic and Transport EngineeringTeaching assistant · CIEM6110 · 2022/23, 2023/24, 2024/25 (Q3)
  • Python ProgrammingTeaching assistant · TIL6022 · 2024/25 (Q1)
  • Master’s student supervisionDaily supervision at TU Delft; co-supervision at UC Berkeley.

Service & leadership

  • IEEE ITSS Benelux ChapterCommunication Chair, 2026–Present · Vice Chair, 2025–2026
  • Peer reviewTransportation Research Parts C and D; Transportation Science; IEEE Transactions on ITS; Transportmetrica A and B.
  • IEEE ICVES 2026Associate Editor
  • Connecting academia & industrySecured IEEE ITSS Young Professionals funding and organized two student/YP events in 2025.Delft networking event ↗Rotterdam industry excursion ↗

05 / ABOUT

Curiosity, on
and off the road.

I am a postdoctoral researcher in Transport & Planning at TU Delft and a visiting scholar at UC Berkeley’s CITRIS and the Banatao Institute, hosted by Prof. Alexandre M. Bayen.

I received my PhD from TU Delft in January 2026, supervised by Prof. Serge P. Hoogendoorn and Dr. Simeon C. Calvert. My dissertation explored driving heterogeneity in traffic-flow theory.

Beyond research, I enjoy snowboarding and skiing, hiking, cycling, and baking. I’m a certified snowboarding instructor, with an enduring interest in psychology and human behavior.

Download my academic CV ↗
A THREAD THROUGH MY WORK
Understanding individual differences.
Making sense of collective behavior.

LET’S CONNECT

Interested in thinking
about traffic together?

[email protected] ↗