Resources

Resources

Mentor Tips & Resources

(Shared during the Application Webinar – April 1)

Tips for Applicants

From Adeel Razi:
    Demonstrate alignment with the core question and clearly articulate your interest in generative modelling, active inference, or understanding adaptive behaviour in neural and/or AI systems.
    Highlight relevant technical skills and experience in Python, data analysis, machine learning, or computational neuroscience will strengthen your application—even if not all are required.
    Show curiosity about interdisciplinary problems as this project sits at the intersection of neuroscience, AI, and dynamical systems, an openness to multiple perspectives is important.
    Emphasise problem-solving and research thinking through prior experience designing experiments, analysing data, or working with models is more valuable than familiarity with specific tools.
    Be comfortable with ambiguity as the project explores an open scientific question without a predefined answer—motivation to engage with uncertainty is key.
    Communicating clearly and concisely as a strong application will explain your interest and relevant experience in a focused and accessible way.
    Interest in neural data or closed-loop systems is a plus and a prior exposure to time-series data, electrophysiology, or interactive systems will be helpful but is not required.

Recommended Papers & Resources

From Winnie Street and Geoff Keeling:
  • A recent short paper from the Google group, closely related to our project:   https://arxiv.org/pdf/2603.28925  
From Aran Nayebi:
  • The Capable Agents paper I was mentioning in my project that it will be based on:  https://arxiv.org/abs/2603.02491 
  • X summary:  https://x.com/aran_nayebi/status/2029234582034272406 
From Ida Momennejad:
  • Here are some of our recent papers using cognitive science and neuroscience methods for algorithmic interpretability of AI behavior and latent space:
    Evaluating cognitive maps and planning in Large Language Models with CogEval. NeurIPS2023  https://arxiv.org/abs/2309.15129  
    A brain-inspired agentic architecture to improve planning with LLMs. Nature Communications 16, 8633.  https://doi.org/10.1038/s41467-025-63804-5  
    Algorithmic Primitives and Compositional Geometry of Reasoning in Language Models.  https://www.arxiv.org/abs/2510.15987#  
    Position: We Need An Algorithmic Understanding of Generative AI. International Conference for Machine Learning, ICML 2025,  https://arxiv.org/abs/2507.07544  
Book recommendation from Michael J Tarr : An advertisement for my colleague’s wonderful book on the science of zombies: Do Zombies Dream of Undead Sheep?: A Neuroscientific View of the Zombie Brain