NeuroAI: Learning the Brain and the Machine Code

NeuroAI: A call for pursuing a general science of brain-inspired and brain-associated computational process

The quest for understanding the nature of learning and intelligence has several origins. Starting from 1950s, scientists and thinkers had the vision to understand the mind from the perspective of machines and computers, and the other direction was found to be very fruitful as well. Conferences had been initiated as vastly cross-disciplinary, but then the communities started to subdivide. In Taiwan, we aim to bring shared-minded people across brain science, machine learning and physical science to a same meeting room for anything we call intelligence.

Practical scope of NeuroAI

NeuroAI has a practical scope of several axes: from basic to applied sciences, as well as from Neuroscience to AI, or vice versa. Rooted in the Taiwanese communities, we are mostly interested in two questions:

(a) How can we frame, interpret and plan neuroscientific research through the lens of "learning" in quantitative or computational disciplines, or machine learning?
(b) How can we apply BNN- or ANN-inspired learning (B and A stand for biological and artificial, respectively) to solve practical problems, e.g., biomedical signal processing, next-generation computing hardware/software, and physically and spatially informed robotics?

The workshop goal

Our major motivation is to bring neuroscientists who subscribe to the potential of AI × Brain to the grand community of the ACML. In practice, experts with diverse backgrounds will be invited to discuss:

1. Brain-inspired learning rules and issues of gradient descent in BNNs
2. Connectome-based intelligence and neuro-engineering perspective of embodied AI
3. Dynamics, mechanics or algorithms in learning systems
4. Brain-inspired computing in hardware and software

The workshop will have a moderated panel discussion to form a digest of the ideas presented with respect to the stated goal, and a lookout for the next steps.

Please submit a 1-page abstract with title, list of authors, formal affiliations, and the abstract body.

We encourage submission of a conference paper of no more than 3 pages. It will not be publicly archived per ACML policies. Participants of the NeuroAI Workshop can access your papers if the first or corresponding authors opt to share them.

Submit Your Paper/Abstract