Abstract:
How do brains and AI turn what they know into multi-step reasoning, planning, and collective problem-solving? I will present a series of empirical and computational studies on how memory and multi-step planning algorithms shape the algorithmic geometry of neural latent spaces and behavior in brains, AI.
Bio: Dr. Ida Momennejad
Ida Momennejad is a Principal Researcher in Machine Learning at Microsoft Research NYC. Her research spans AI, cognitive neuroscience, and philosophy, studying underlying algorithms of memory and reasoning in brains and AI: how memories are structured to support both remembering and foresight, how brains and AI acquire models of the world and use them for reasoning and planning, and which multi-agent architectures best support collective memory and hierarchical innovation. Philosophically, she is moving beyond computational functionalism, regarding mathematical abstractions of living intelligence as tools for understanding rather than minds.