What AI Can Do for Neuroscience: Understanding How the Brain Represents Word Meanings (2023)
To show what AI can do for neuroscience, this chapter presents a case study from a collaboration between neuroscience research, its tools (functional magnetic resonance imaging – fMRI) and theories (Concept Attribute Representation model - CAR model), and an AI approach (context-dependent meaning representation in the brain – CEREBRA model). Such interaction produced new opportunities to allow researchers to gain insights and validate some hypotheses about the functioning of the brain (within the language domain) and delivered a unique class of dynamic word representations (based on the way word meanings are represented in the brain) that may improve current natural language processing (NLP) systems such as Siri, Google, and Alexa, by dynamically adapting their representations to fit context.
In Manuel Cebral-Loureda, Elvira G. Rincon-Flores, and Gildardo Sanchez-Ante, editors, What AI Can Do: Strengths and Limitations of Artificial Intelligence, 401-417, 2023. CRC Press.

Nora E. Aguirre-Celis Ph.D. Alumni naguirre [at] cs utexas edu
Risto Miikkulainen Faculty risto [at] cs utexas edu