Latest News & Reviews

Royal Institution Lecture on Youtube – 68K views after 4 weeks!

July 31, 2021 Events News

The Royal Institution hosted a one and a half hour live lecture and discussion on The Self-Assembling Brain and the cross-talk of biological and artificial neural network research. Thanks to 800 participants for signing up, watching it live or over the next week via the Royal Institution video link.

The lecture is online since July 2 on Youtube. Thanks for 2.7K likes in the first 4 weeks!

Here are just a few of the comments and fabulous questions from the discussion:

“From someone who played gameoflife at PARC in 1980, I like your intelligence in approaching these very difficult problems. Thank you for a fascinating talk. I learned a lot and history too.”

“One of the most interesting talks I’ve ever heard.”

“Does evolutionary programming have applications in cryptography? they seem to have similarities to one way functions”

“Are there people working in AI to make different types of Artificial Intelligence as different as we are from bees?”

“In your model, genome, time and energy is required for the development of a neural network, but this is similar for other organs just on a shorter time scale. So what makes a neural network different from any other organ, is it the time?”

“How successful do you think Elon Musk’s Neuralink company will have in making a breakthrough in cracking neural networks and making it practically usable in medicine and every day life?”

Biologically speaking, I imagine how mind-blowing it would be to reproduce a human brain in a computer, but do you think this is computationally useful? I mean, instead of having a copy of a human brain I think it would be better for us to create something different that can give us a different point of view on a variety of tasks, like architecture or art?”

“What are your views on developing a relationship between human brain and animal brain if it varies from worms or insects or anything ?”

“Current NN and DL in most cases are machine supervised learning (i.e. these eat big data, need it), not quite AI I would argue. Once unsupervised learning becomes meaningful, self-learning is a next step, still no AI I would argue. Reasoning is the still missing component, agree?”

“then ai are not smarter than us they are just faster than us at learning”

“**Could finding an ensemble learning technique (a method of using multiple models together) using natural selection be akin to encoding a genome using growth over time?**”

“Can you imagine what type of evolutionary improvement to human intelligence could eventually develop?”

“We’re quite far away. You’re right. Our priorities are quite different at the moment. We can take a look at GPT-3. “

“How dangerous is the lack of caring about the nature of the neural network process in allowing bias to be taught or learned?  If you teach data with only white people, what happens when you show the network a person of color?”

“Wonderful talk sir, Really looking forward to at least make a contribution in the near future”

“Thanking everyone as well for asking such insightful questions. Really opened my mind”

Many more interesting (and critical!) comments on Youtube.