Learning in artificial, real, and social brains
We're all familiar with the superior ability of AI systems on very specific problems in recent years. For my generation, chess was the benchmark for the rise of machines. In yours, it may be your favourite video game, or possibly more worrying, the ability of machines to take over yet another avenue of employment, such as diagnosis or chip design. One of the big distinctions remains learnability. If you had a colleague who needed to sit through a class a few million times to learn a concept, you would not be very impressed. I'll sketch how human and animal brains learn things, and although some of the core concepts - such as reinforcement learning – have been taken over to AI, we now have many clues about why real brains learn so efficiently. Ultimately our human ability to learn as individuals and as societies is more limited by what new ideas we choose to explore, than by intrinsic capabilities. We had better boost it because the machines are coming.
Dr Upinder Singh Bhalla, an Indian computational neuroscientist and Professor at NCBS (National Center for Biological Sciences) of the Tata Institute of Fundamental Research; Elected Fellow of the Indian Academy of Sciences and the Indian National Science Academy; Recipient of the Shanti Swaroop Bhatnagar Prize for Science and Technology by CSIR.
Dr Bhalla is known for his studies on neuronal and synaptic signalling in memory and olfactory coding using computational and experimental methods. Computational neuroscience is an emerging domain with lots of research being carried out, with vast practical implications.
Date: 02nd April 2021.
Time: 12 pm to 1 pm (IST)
Contact : Arnab - +91 8240115799
Ramachandra - +91 9487413092