How Neuroscience Shaped Modern AI

• ​A lot of what powers modern AI actually traces back to ideas from neuroscience and some of them are decades old.

• ​This talk looks at how concepts like neural architectures, learning dynamics, and scaling laws came together to make today's AI systems possible.

• ​We'll walk through the journey from early neural models to deep networks and transformers, and where the parallels with the brain show up.

• ​We'll also dig into what this convergence might tell us about where AI is headed next.

• ​And finally, some open questions where neuroscience and ML might keep learning from each other.

​About the speaker:
Arvind Saraf is a computer engineer, scientist, entrepreneur, and systems builder. He previously co-founded a Ratan Tata–backed social impact venture and an InfoEdge backed startup, served as VP of Engineering at Drishti, later acquired by Apple, and currently works on large-scale AI systems as part of Microsoft’s Turing team.
linkedin.com/in/arvind-saraf

​
To attend online:

• ​⁠Add to calendar: https://bit.ly/4pNTtaD

• ​Gmeet link: https://meet.google.com/cwb-zoro-die?hs=122&authuser=0

​Pre-read:
Neuroscience basics

1. ​“Computational cognitive neuroscience” (book), O’Reilly etal, esp chapters 3,4,5

2. ​“Introduction to Neural Computation”, MIT OCW - https://ocw.mit.edu/courses/9-40-introduction-to-neural-computation-spring-2018/resources/20/ 

3. ​“Neural mechanisms of selective visual attention”, Desimone etal

4. ​“Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects”, Rao etal

5. ​“A Brief History of Intelligence: Why the Evolution of the Brain Holds the Key to the Future of AI”, Bennett

6. ​Alignment studies:

   1. ​“Representational similarity analysis – connecting the branches of systems neuroscience”, Kriegeskorte etal - https://www.frontiersin.org/journals/systems-neuroscience/articles/10.3389/neuro.06.004.2008/full 

   2. ​“Similarity of Neural Network Representations Revisited”, Kornblith etal - https://arxiv.org/abs/1905.00414 

   3. ​“Artificial Neural Networks for Neuroscientists”, Yang etal

   4. ​“Performance-optimized hierarchical models predict neural responses in higher visual cortex”, Yamins etal - https://www.pnas.org/doi/epdf/10.1073/pnas.1403112111 

   5. ​“Emergence of human-like attention and distinct head clusters in self-supervised vision transformers: A comparative eye-tracking study”  - Yamamoto etal

   6. ​“Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning”, Lotter etal

7. ​Sutton’s essays:

   1. ​“Welcome to the Era of experience”, Sutton etal - https://compcogneuro.org https://storage.googleapis.com/deepmind-media/Era-of-Experience%20/The%20Era%20of%20Experience%20Paper.pdf 

   2. ​“The Bitter Lesson”, Sutton etal