How AI's Storming the Fashion Industry - Ep. 61

Nature's Lessons

Costa Colbert emphasizes the importance of learning from nature, predicting that future AI advancements will draw heavily from biological systems. He notes that although some believe we've learned all we need from nature, there's still much that biological brains do better than computers. Colbert mentions backpropagation, a key algorithm in AI, and how its biological plausibility is still debated.

I think we'll find that there's still many things that biological brains do much, much better than computers.

--- Costa Colbert

He believes that the debate on whether the brain uses backpropagation will continue, with both neuroscience and AI communities contributing to the discussion 1.

   

Bridging the Gap

The integration of neuroscience and AI has evolved significantly over the years. Costa Colbert explains that while neuroscience was traditionally a biological field, the rise of deep learning has brought computational questions to the forefront. He highlights the historical disconnect between systems neuroscience and computational neuroscience, noting that younger researchers are now bridging this gap.

The computational neuroscience was always pushed to the fringe a certain amount. And I think now maybe that will start to change.

--- Costa Colbert

Colbert believes that this integration will lead to exciting advancements in both fields 2.

   

Journey into AI

Reflecting on his journey, Costa Colbert shares how his interest in AI began in the 1980s with a book on Lisp, a programming language for symbolic AI. His fascination with the biological basis of learning and memory led him to focus on understanding neurons and synapses. Despite the initial skepticism, his work in deep learning has become increasingly relevant to modern AI applications.

The first time I got interested in artificial intelligence was 1980, and I had a summer job that was very boring.

--- Costa Colbert

Colbert's diverse background in neuroscience, electrical engineering, and computer science has uniquely positioned him to contribute to the evolving landscape of AI 3.