Activities & Schedule

Symposium Activities

The symposium will engage attendees through a variety of sessions designed to encourage both learning and dialogue:

  • Invited speakers: Distinguished scholars in human-AI collaboration will provide keynote talks, setting the context for exploring trust, transparency, and responsible design across application areas.
  • Oral presentations and poster sessions: Accepted research papers and posters will showcase new insights and innovations, focusing on enhancing communication, transparency, and ethical standards in human-AI interactions. Neuro-symbolic approaches and applications of LLMs, VR/AR, and NLP
    in collaborative settings are welcome.
  • Tutorials: Tutorials will offer hands-on instruction in key technologies that support human-AI collaboration, including:
    • Explainable AI Techniques: Sessions on using interpretability methods like SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) for understanding AI outputs.
    • Human-Centred LLM and Neuro-Symbolic Design: Instruction on prompt engineering, symbolic rule integration, and ethical design strategies to align AI responses with user expectations.
    • Collaborative Machine Learning Platforms: Tutorials on platforms like TensorFlow Federated for privacy-preserving, decentralised collaborative learning.
  • Panel session: A focused discussion on the “Impact of diversity of human and computational intelligence on the design of hybrid intelligent systems” will allow panellists to address critical questions such as:
    • How can the diversity in human cognitive abilities and AI models be laveraged to enhance the performance and adaptability of hybrid intelligent systems?
    • What are the challenges and best practices in integrating diverse human inputs and AI capabilities within hybrid systems?
    • How can we ensure that hybrid intelligent systems are designed ethically, respecting diverse human perspectives while mitigating biases in AI?
  • Workshops: Intensive workshops will explore specific methodologies and technologies for human-AI collaboration. Topics include:
    • Bias Mitigation and Fairness in AI Models: Approaches for identifying and reducing bias in AI outputs, with tools and best practices for promoting fairness.
    • Explainability Tools for User Trust: Techniques for applying explainability methods, including symbolic rule-based explanations, to improve user confidence in AI-driven applications.
    • RecSim and Neuro-Symbolic Approaches for User Modelling: This workshop will cover Google’s RecSim for creating simulations that assess and optimise recommendation systems alongside neuro-symbolic approaches to model user behaviour.
    • Human Augmentation Technologies: A session focusing on AI-driven advancements in human augmentation, such as cognitive prosthetics, AI-assisted diagnostics, and performance enhancement tools.
  • Roundtable Discussion: Exploring the benefits and challenges of establishing an international network for Human-AI collaboration. A proposal will be presented to attendees for feedback and possible adoption.

Schedule

Day 1

Morning (Opening & paper presentations)

  • 8:00 – 8:45: Registration and Welcome
  • 9:00 – 10:00: Keynote on human-AI collaboration – Vision, Challenges, and Opportunities
  • 10:00 – 11:20: Oral session 1: Human-AI interaction
    Four selected papers (20 minutes each) focusing on foundational concepts of human-AI collaboration, user-centered design, and building trust in AI systems.
  • 11:20 – 11:45: Networking coffee break
  • 11:45 – 13:15: Tutorial session 1: Explainable AI techniques
    Hands-on tutorial on explainability methods like SHAP and LIME to improve transparency and interpretability of AI systems, followed by Q&A.

Afternoon (Oral presentations and tutorials)

  • 14:00 – 15:20: Oral session 2: Explainability and interpretability
    Four selected papers (20 minutes each) discussing techniques for improving explainability in AI, including symbolic knowledge integration, model transparency, and ethical considerations in interpretability.
  • 15:20 – 15:45: Networking coffee break
  • 15:45 – 17:15: Tutorial session 2: Human-centered LLM and neuro-symbolic design
    Tutorial on incorporating symbolic knowledge with neural methods, covering prompt engineering, symbolic rule integration,
    and ethical design strategies.

Day 2

Morning (Oral session and poster)

  • 9:00 – 10:00: Keynote on Explainable AI
  • 10:00 – 11:20: Oral session 3: Applications of human-AI collaboration
    Four selected papers (20 minutes each) on real-world applications of AI in healthcare, education, and creative industries, with a focus on domain-specific challenges and solutions for human-AI interaction.
  • 11:20 – 11:45: Networking coffee break
  • 11:45 – 13:15: Poster session and networking
    Poster session featuring research on NLP applications, user adaptation, cognitive aspects in collaborative AI, and practical challenges.

Afternoon (Workshop and social event)

  • 14:00 – 15:20: Workshop 1: Bias mitigation and fairness in AI models
    Workshop covering practical tools and strategies to identify and mitigate bias in collaborative AI applications.
  • 15:20 – 15:45: Networking coffee break
  • 16:00 – 18:00: Social event – Visiting the future meusium

Day 3

Morning (Panel session, Oral session, and Workshop)

  • 9:00 – 10:30: Workshop 2: RecSim and user modeling
    Workshop on Google’s RecSim for recommendation system simulation, aiming at enhanced user interaction modeling.
  • 10:35 – 11:00: Coffee break
  • 11:00 – 12:30: Panel session
    Theme: “Impact of diversity of human and computational intelligence on hybrid intelligent systems.”
    Panellists discuss the diversity of human and computational intelligence and impact on hybrid intelligent systems.
  • 13:00 – 2:00: Roundtable discussion:
    Establishing an international network on ”Human-AI collaboration.”