The second InfoCog@NeurIPS workshop aims to further the understanding of information flow in cognitive processes and neural networks models of cognition by bringing together researchers from multiple disciplines, including machine learning, cognitive science, neuroscience, and information theory. This is the second workshop of this kind in NeurIPS. This year the workshop will continue last year’s community-building momentum and will additionally focus on advancing two main goals: First, providing a fruitful platform for discussing formulations of storage and processing of information in human and artificial cognitive systems, via information-theoretic measures. Specifically, we aim to further integrate information theory researchers in such discussions, deepening the mathematical knowledge and tools shared across disciplines. Second, facilitating research on addressing the challenges of computing information theoretic quantities, specifically for the needs of cognitive science and machine learning.

We invite short submissions (up to 4 pages – NeurIPS format) that present original work related to information theory and cognitive systems, and particularly encourage submissions of collaborative interdisciplinary work. Examples of specific topics of interest include but are not limited to:

• Novel information-theoretic approaches to cognitive functions (e.g., perception, decision making, language, social reasoning, etc.)
• Method and approaches for the validation of information theoretic formalisms in human and artificial cognition
• Novel methods and approaches for the computation and/or estimation of information theoretic quantities and their applications to human and artificial cognition
• Challenges and limitations of the use of information theory in studying cognitive systems
• Application of information theory to training human-aligned artificial agents, i.e., agents that can better communicate and cooperate with humans

All accepted papers will be presented in a poster session during the workshop, and a number of papers will also be selected for an oral presentation. We expect to have limited resources to offer financial aid to selected participants. More details will follow upon acceptance.

https://sites.google.com/view/infocog-neurips-2023/call-for-papers