Conference Program on January 21, 2026 - Room: Garnet 215

Making Embodied AI Reliable with Testing and Formal Verification

• 08:30 – 08:45 | Opening Remarks
• 08:45 – 09:30 | Invited Talk
  Speaker: Prof. Guy Van Den Broeck (UCLA)
  Title: Symbolic Reasoning in the Age of Large Language Models
• 09:30 – 10:30 | Technical Session 1 – Embodied AI, Reinforcement Learning, World Models & Multi-Robot Safety (4 Papers)

  Each paper: 10-minute presentation + 5-minute Q&A

  1. Inferring Causal Graph Temporal Logic Formulas to Expedite Reinforcement Learning in Temporally Extended Tasks
     Hadi Partovi Aria and Zhe Xu
  2. Towards a Feedback-driven Adaptive Embodied Environment Generator for Training Robotic Agents
     Teresa Yeo, Dulaj Weerakoon, Dulanga Weerakoon and Archan Misra
  3. Explicit World Models for Reliable Human-Robot Collaboration
     Kenneth Kwok, Basura Fernando, Qianli Xu, Vigneshwaran Subbaraju, Dongkyu Choi and Boon Kiat Quek
  4. Robust Evacuation for Multi-Drone Failure in Drone Light Shows
     Minhyuk Park, Aloysius K. Mok and Tsz-Chiu Au
• 10:30 – 11:00 | Coffee Break
• 11:00 – 12:15 | Technical Session 2 – Verification, Certification, Program Analysis & LLM Safety (5 Papers)

  Each paper: 10-minute presentation + 5-minute Q&A

  1. Formal Safety Guarantees for Autonomous Vehicles using Barrier Certificates
     Oumaima Barhoumi, Mohamed H Zaki and Sofiène Tahar
  2. A New Strategy for Verifying Reach-Avoid Specifications in Neural Feedback Systems
     Sam Akinwande, Sydney Katz, Mykel Kochenderfer and Clark Barrett
  3. Toward Maturity-Based Certification of Embodied AI: Quantifying Trustworthiness Through Measurement Mechanisms
     Michael Darling, Alan Hesu, Michael Mardikes, Brian McGuigan and Reed Milewicz
  4. BALI: Branch-Aware Loop Invariant Inference with Large Language Models
     Mingxiu Wang, Jiawei Wang and Xiao Cheng
  5. A Unified Framework for Jailbreak Attacks on Large Language Models
     Xiaobing Sun and Liangli Zhen
• 12:15 – 12:45 | Poster Session
  (All accepted papers may optionally bring posters.)
• 12:45 – 14:00 | Lunch Break
• 14:00 – 14:45 | Panel Discussion

  Moderator: Xi Zheng

  Theme: Certification Challenges for AI-Enabled Autonomy: What Can Neurosymbolic Methods Enable?

  Panellists:

  • Guy Van Den Broeck (UCLA)
  • Dinesh Manocha (UMCP)
  • Taylor T. Johnson (Vanderbilt University)
• 14:45 – 15:30 | Breakout Discussions
  Participants split into three focused groups:
  1. Scenario-Based Testing and LLM-Guided Specification Mining (led by Xi Zheng)
  2. Compositional Verification and Modular Reasoning for Embodied AI (led by Guy Van Den Broeck)
  3. Robustness Under Uncertainty, Sensing Noise, and Human–Robot Interaction (led by Archan Misra)

  Each group produces 2–3 actionable recommendations for the roadmap.

• 15:30 – 16:00 | Coffee Break
• 16:00 – 16:30 | Breakout Report-Back Session
  Each group presents their key insights and open research questions.
• 16:30 – 16:50 | Roadmap Consolidation Discussion
  Collective shaping of a community roadmap for reliable embodied AI, covering:
  • Neurosymbolic architectures
  • Testing and scenario generation
  • Specification mining
  • Verification and certification
  • Robustness and uncertainty management
• 16:50 – 17:00 | Closing Remarks

  Plans for:

  • Community follow-up
  • Special issues / joint whitepaper
  • Future AAAI Bridge events and co-located workshops