originally published on linkedIn: https://www.linkedin.com/posts/amprather_ai-activity-7109242104911941632-vBmy
Director, Robotics & Autonomous Systems Program at ASTM International
As we continue to ponder where #AI is going, it is probably a good idea to look at our own brains and how we developed our own intelligence. This was one of the reasons I picked up a copy of “𝐓𝐡𝐞 𝐒𝐞𝐥𝐟-𝐀𝐬𝐬𝐞𝐦𝐛𝐥𝐢𝐧𝐠 𝐁𝐫𝐚𝐢𝐧: 𝐇𝐨𝐰 𝐍𝐞𝐮𝐫𝐚𝐥 𝐍𝐞𝐭𝐰𝐨𝐫𝐤𝐬 𝐆𝐫𝐨𝐰 𝐒𝐦𝐚𝐫𝐭𝐞𝐫” by 𝐏𝐞𝐭𝐞𝐫 𝐑𝐨𝐛𝐢𝐧 𝐇𝐢𝐞𝐬𝐢𝐧𝐠𝐞𝐫. The premise is somewhat straight forward. How does a neural neural network become a brain? Neurobiologists explore how nature created our own brains, but computer scientists look at how technology can achieve it as well in the form of Artificial Intelligence. For biology, brains grow based on genes and are already smart before they even start learning (that is mind blowing within itself). By contrast, artificial neural networks are designed, not grown, and start out with random connections and through these connections get smarter through learning alone. Hiesinger starts the book off with telling the history of the two fields – neurobiology and artificial intelligence. This probably could have been its own text book alone. However, after that it pivots to a group of scientists from the various fields, including a robotics engineer, having ongoing conversations and arguments between the chapters. To say they agree on many things is an understatement. However, the conversations they have are very timely and up-to-date. This may be a problem for the book in 5 to 10 years. This book was a fun read, but not a light read. I had to stop in some chapters to digest some of the concepts before moving on. The topics get very heavy as the book continues on through the complexities on both the biology and AI sides. I think the biggest impact that came from the book is that AI can’t create overnight what nature has been constructing through centuries of evolution in not only humans, but other creatures. AI engineers need to understand this concept, because it may help form the next evolutionary steps that need to take place in AI. Acknowledging the significance of developmental biology and evolution when tackling learning with AI is going to be important. This is definitely, a perplexing, controversial and thrilling concept that could mark the beginning of new approaches in AI. The question will be can the biologists and technologists agree to what those approachs should be. I finished this book with even more questions than I had coming in, but that is probably a good sign of how this book made me rethink some of my previous thoughts on how both my own brain developed and how an AI brain could. Pick up a copy here: https://lnkd.in/eUkp_Bjq