AI Frontiers: Rethinking intelligence with Ashley Llorens and Ida Momennejad

Powerful large-scale AI models like GPT-4 are showing dramatic improvements in reasoning, problem-solving, and language capabilities. This marks a phase change for artificial intelligence—and a signal of accelerating progress to come.In this Microsoft Research Podcast series, AI scientist and engineer Ashley Llorens hosts conversations with his collaborators and colleagues about what these models—and the models that will come next—mean for our approach to creating, understanding, and deploying AI, its applications in areas such as health care and education, and its potential to benefit humanity.This episode features Principal Researcher Ida Momennejad. Momennejad is applying her expertise in cognitive neuroscience and computer science to better understand—and extend—AI capabilities, particularly when it comes to multistep reasoning and short- and long-term planning. Llorens and Momennejad discuss the notion of general intelligence in both humans and machines; how Momennejad and colleagues leveraged prior research into the cognition of people and rats to create prompts for evaluating large language models; and the case for the development of a “prefrontal cortex” for AI.Learn more:* AI and Microsoft Research (https://www.microsoft.com/en-us/research/focus-area/ai-and-microsoft-research/) | Focus Area* Evaluating Cognitive Maps and Planning in Large Language Models with CogEval (https://www.microsoft.com/en-us/research/publication/evaluating-cognitive-maps-in-large-language-models-with-cogeval-no-emergent-planning/) | Publication, October 2023* Imitating Human Behaviour with Diffusion Models (https://www.microsoft.com/en-us/research/publication/imitating-human-behaviour-with-diffusion-models/) | Publication, May 2023* Navigates Like Me: Understanding How People Evaluate Human-Like AI in Video Games (https://www.microsoft.com/en-us/research/publication/navigates-like-me-understanding-how-people-evaluate-human-like-ai-in-video-games/) | Publication, April 2023* Navigation Turing Test (NTT): Learning to Evaluate Human-Like Navigation (https://www.microsoft.com/en-us/research/publication/navigation-turing-test-ntt-learning-to-evaluate-human-like-navigation/) | Publication, July 2021* Predictive Representations in Hippocampal and Prefrontal Hierarchies (https://www.microsoft.com/en-us/research/publication/predictive-representations-in-hippocampal-and-prefrontal-hierarchies/) | Publication, January 2022* The successor representation in human reinforcement learning (https://www.microsoft.com/en-us/research/publication/the-successor-representation-in-human-reinforcement-learning/) | Publication, September 2017* Encoding of Prospective Tasks in the Human Prefrontal Cortex under Varying Task Loads (https://www.microsoft.com/en-us/research/publication/encoding-of-prospective-tasks-in-the-human-prefrontal-cortex-under-varying-task-loads/) | Publication, October 2013