Schedule

08:00 – Michael Buice, Learning dynamics of deep networks with multiple pathways

08:20 – John Kruper (Ariel Rokem lab), Effects of aging on tissue properties of the optic radiations, and how this differs with glaucoma

08:40 – Tristan Glatard, LivingPark: open evaluations of Parkinson's Disease MRI measures

09:00 – Tim Murphy, Automated methods to model brain and behavior connections in mouse cortex: a focus on stroke

09:20 – Elliott Abe (Cris Niell lab), Joint coding of visual input and eye/head position in V1 of freely moving mice

09:40 – Daniela Witten, Identifying subpopulations of neurons without double-dipping

10:00 - Break

10:20 – Joern Davidsen, Chimera states & the critical brain hypothesis

10:40 – Beth Buffalo, Visual exploration signals in the monkey hippocampus

11:00 – Bing Brunton, Tracking turbulent plumes with deep reinforcement learning

11:20 – Santiago Jaramillo, Learning from combinations of active training and passive exposure to sounds

11:40 – Nick Steinmetz, Distributed coding for perception, action, and cognition in the mouse brain

12:00 - Lunch

01:00 – Wilten Nicola, Supervised Learning Rules in the Hypothalamus Mediate a Plastic Response to Stressors

01:20 - Adrienne Fairhall

01:40 – Kameron Harris, Network geometry for sensing and learning

02:00 – Break (IN-BIC planning meeting)

3:00 – Shanahan Foundation Fellows

5:00 – Plenary Talk: Blaise Agüera y Arcas, AI: where did it come from, what is it now and where is it going?

6:00 – Reception

08:15 - Introduction, Frederic Jung, Consul general of France in San Francisco and Mireille Guyader, Counselor for Science and Technology at the Embassy of France in the United States

08:30 – Frederic Alexandre, Deciphering the Biological Basis of Cognitive Control

09:00 – Blake Richards, Contrastive introspection for brain-like credit assignment in reinforcement learning

09:30 – Ida Mommenajad, A rubric for Human-like Artificial Agents and NeuroAI

10:00 - Break

10:30 – Eric Shea-Brown, Assigning credit through the "other” connectome

11:00 – Lia Papadopoulos (Mazzucato Lab), Metastable circuit dynamics explain optimal coding of auditory stimuli at moderate arousals

11:30 – Lunch and Poster Session

01:00 - Jeremiah Cohen, Norepinephrine neurons drive reinforcement learning

01:30 – Boris Gutkin, Needing, Tasting, Wanting: A Homeostatic Framework for Reinforcement Learning

02:00 - Anne Collins, Insights into the computations supporting intelligent human behavior

02:30 - Aaron Gruber, Use of schematic knowledge in reinforcement learning tasks

03:00 - Break

03:30 - Reinforcement learning discussion

04:30 - Rajesh Rao, Dynamic and Active Predictive Coding: New Approaches to Understanding Cortical Function

06:00 – Reception at Agua Verde (1303 NE Boat St, Seattle, WA 98105)

09:00 – Cornelia Fermuller, Emulating the motion pathway

09:30 – Jean-Baptiste Masson, Towards understanding the underlying principles of small biological neural network design

10:00 - Guillaume Lajoie, How deep learning theory can inform and benefit from brain structure and learning dynamics

10:30 – Jean Daunizeau, Synaptic plasticity in the orbitofrontal cortex explains how risk attitude adapts to the range of risk prospect

11:00 – Lunch and Poster Session

12:30 – Amy Orsborn, Measuring, modeling and shaping neural plasticity in brain-machine interfaces

01:00 – Stefan Mihalas, Computing with a mess: how complex and heterogeneous components help network computation

01:30 – break

02:00 - Hannaneh Hajishirzi, Toward Robust and Knowledge-Rich Natural Language Processing

02:30 - Eli Shlizerman, Unsupervised and semi-supervised learning for interpreting and connecting behavior with brain activity

03:00 - Corinne Teeter, Cognitive and Emerging computing at Sandia

03:30 - Break

04:00 - Industry research panel with

Julie Harris (Cajal Neuroscience)

Babak Parviz (Amazon)

Philip Sabes (Starfish Neuroscience)

Corinne Teeter (Sandia)

05:30 - Trainee Mixer at Big Time Brewery (4133 University Way NE, Seattle, WA 98105)

Advances and challenges in AI/ML for neurotechnologies

Neural interfacing technologies, such as brain-computer interfaces and neural stimulation paradigms, are under rapid development. Recent years have been witness to astounding progress in hardware, with ever increasing sensitivity, channel count, and general bandwidth for both invasive and non-invasive neural interfacing devices. While these technologies have unprecedented potential clinical and scientific promise, advances in hardware for measuring and manipulating the nervous system have outpaced our development of algorithms to use them. here remain important challenges in deriving a robust and scalable code with which we can decode information into and encode information from the nervous system, using high-throughput devices.

This workshop aims at exploring the algorithmic challenges facing neural interfacing in the near future. We will explore the unique difficulties of AI and ML development for neurotechnologies, including: noisy and variable signals, non-stationarity and co-adaptability of brains and machines interacting, distributed information decoding and encoding at the neural population level, and more.

The goal is to promote multidisciplinary discussions between experimentalists and AI/ML practitioners, as well as industry and academia, to identify promising areas of development for scalable and robust neural interfacing algorithms.

Schedule

08:30 – Opening remarks

08:45- Yann Le Cun

09:30 – Eb Fetz, Volitional control of neural activity

10:30 - Break

11:00 – Matt Golub

11:30 – Matt Perich

12:00 – Lunch

01:00 – Anne Draelos, Real-time modeling with adaptive interventions for high-dimensional neural data

01:30 – Maryam Shanechi

02:00 - James Murray, Distinguishing learning rules with brain-machine interfaces

02:30 - Break

03:00 - Emily Mugler

03:30 - Philip Sabes, AI for brain interfaces: challenges and opportunities

04:30 - Panel

05:00 - Reception