Following our series of CIMCYC Sessions, we are pleased to announce our next AI, Mind and Brain session on Wednesday, 18th June. As in other sessions, we will have an unguided discussion, so do come with ideas for discussion (5 minutes of thinking time). Find below the full reference of the paper, the link to a podcast summary of the paper and a couple of questions to inspire discussion (via Notebook LM and ChatGPT).
Paper: Dujmovic, M., Bowers, J., Adolfi, F. and Malhotra, G. (2024). Inferring DNN-Brain Alignment using Representational Similarity Analyses can be Problematic. ICLR 2024 Workshop on Representational Alignment. https://openreview.net/pdf?id=dSEwiAENTS
Podcast summary: bit.ly/45bAvEg
Kickstarting questions:
-
The proposed paper highlights the tension between DNNs having high RSA with brain data, but failing many standard psychology tests. How does the problem of ‘second-order confounds’ help explain this discrepancy and challenge what we mean by ‘DNN-brain alignment’?
-
If a high RSA score between two systems does not guarantee they encode the same stimulus features, what specific type of similarity can we still infer from their representational geometries, based on the views discussed on the proposed paper?
No prior experience in AI is needed –just curiosity and a willingness to discuss ideas. Save the date!
*Please reach out to @email before the meeting in case you need confirmation for your attendance.
Date: Wednesday, 18th June 2025
Location: Seminario 4 - CIMCYC
Time: 10.00 am
Join us online too!: meet.google.com/sgz-quvr-rma
Upcoming CIMCYC Sessions
-
Wednesday, 25th June: EEG+ Q&A
Instead of a traditional talk, we will host a Q&A session focused on EEG+. To make sure there is a good range of questions, we have created a short Google Form. Please fill it out: bit.ly/45dPZrq We’re looking forward to your questions!
-
Thursday, 3rd July: Joint event at Carmen de la Victoria (coffee break & snacks included for registered attendees)
Francisco Jesús Martínez-Murcia - Introduction to autoencoders and EEG + Q&A: ‘What can AI add to our use of EEG to understand cognition?’