And that's a wrap!

Thank you to @gtsciences.bsky.social and all of today's #GTFrontiers speakers for a thrilling day of neuroscience and AI. #Neuroscience #AI

@apurvaratan.bsky.social aims to apply how intelligence is organized in the brain to improve AI.

For example, visual information is highly structured and topographically organized. The same can't be said about AI — so how do we introduce that topography? #GTFrontiers

Celikel posits that it's all about connections — particularly quantifying how neurons are connected in the brain.

"What we need is an engine that can infer and control, and in order to do that we’re going to need to reverse-engineer the brain." #GTFrontiers #AI #Neuroscience

These varying cell types and their connectivity may be a key component to how our brain filters out noise — and incorporating different cell types into computational models may improve their stability.

More here 🧠: neuro.gatech.edu/sharper-imag... #AI #Neuroscience #GTFrontiers

We've talked about what AI can tell us about the brain, but what can the brain tell us about AI?

Next we have a panel on Advancing AI with Cognitive Science and Neuroscience, featuring GT faculty Hannah Choi, Tansu Celikel, Nabil Imam, and @apurvaratan.bsky.social. #GTFrontiers

We expect AI models to quickly make decisions and predictions.

How do these things occur in the brain (do neurons really make predictions?), and how can we map that computationally with AI? #Neuroscience #ArtificialIntelligenec #AI #GTFrontiers

The final keynote of the day comes (virtually) from Hal Greenwald, where he will share brain-inspired computing research conducted at the Air Force Research Laboratory. #GTFrontiers

Our next session starts with remarks from Tim Lieuwen, Executive Vice for @gtresearchnews.bsky.social .

Lieuwen highlighted Georgia Tech's dedication to advancing neuroscience through new academic programs and the upcoming Institute for Neuroscience, Neurotechnology, and Society. #GTFrontiers

Finally, we hear from Winjing Liao.

Larger AI models and bigger datasets generally lead to better performance. Liao aims to understand why by exploiting low-dimensional structures in data. #GTFrontiers #Neuroscience #LLMs #ChatGPT

There has been an explosion of data from the brain in recent years, and we can now record from more neurons than ever.

Sederberg wants to know we get from detailed data like recordings from groups of neurons to larger cognitive functions. #GTFrontiers #Neuroscience #AI

Predicting the future is key to our survival, and updating our predictions is how we learn.

Najafi studies how different areas of the brain perform this predictive processing and how those areas interact. Can AI help us undercover the answers? #GTFrontiers #InternalModels #Neuroscience

Now, we'll have a panel on Advancing Neuroscience with AI, featuring Georgia Tech faculty Farzaneh Najafi, Audrey Sederberg, @anqiwu.bsky.social, and Wenjing Liao. #GTFrontiers #Neuroscience #AI #ArtificialIntelligence @gtsciences.bsky.social

We can measure our environment and human behavior effectively, but understanding how our brains process this information to make decisions is challenging. By adapting AI models to mimic human behavior, AI could help us understand brain structure and decision-making processes. #GTFrontiers #Neuro #AI

Next, we’ll have a panel on Advancing Cognitive Science with AI featuring Georgia Tech faculty @neuranna.bsky.social, @dobyrahnev.bsky.social, @crozell.bsky.social, and Sashank Varma.

Can AI be used to reveal the internal processing of our minds, which is famously difficult to measure? #GTFrontiers

Momennejad studies reinforcement learning, where agents learn by interacting with their environment, receiving rewards or penalties based on their actions.

How can we computationally map memories of these rewards to optimize choices, and is this similar to how the brain maps memories? #GTFrontiers

The first talk comes from Ida Momennejad (@neuroai.bsky.social), a senior researcher in reinforcement learning at Microsoft Research.

How do we use our memories to make behavioral decisions, what does that look like in the brain? Can we replicate that computationally? #GTFrontiers #AI #Neuroscience