Yann LeCun - Gen AI Winter School, Objective Driven AI.
Вставка
- Опубліковано 14 кві 2024
- 🔥 Breaking! 🌟 Don't miss out on this rare gem! The moment you've been waiting for is finally here! The video of Yann LeCun, the fifth and final conference from the first day of our Winter School on Generative AI at Emines-UM6P is now available!
🎥 Join us in celebrating the insightful presentation by Professor Yann LeCun, the distinguished recipient of the 2018 Turing Award, also known as the Nobel Prize of Computer Science. Currently serving as the VP & Chief AI Scientist at Meta and holding the prestigious Silver Professorship at New York University, Professor LeCun recently gained attention with the launch of LLaMA, a groundbreaking AI model developed by Meta. Despite its smaller size compared to competitors, LLaMA is designed for greater efficiency and resource-friendliness, making a significant contribution to the field of AI.
💡 In his thought-provoking conference on 𝗢𝗯𝗷𝗲𝗰𝘁𝗶𝘃𝗲 𝗗𝗿𝗶𝘃𝗲𝗻 𝗔𝗜, Professor LeCun explored the path towards a system capable of learning, remembering, reasoning, planning, having common sense, all while being directed and safe.
🏆 Let's also extend our heartfelt gratitude to Professor LeCun for presiding over the hackathon proposed during the winter school and for awarding prizes alongside Amal, Eric, and Mohammed.
🔗 Watch the video here.
🌟 Don't miss this opportunity to learn from one of the foremost experts in AI and to gain insights into the future of the field!
Organized by: Prof. Mohammed El Rhabi, Prof. Eric Moulines, and Prof. Nicolas Cheimanoff.
Join us in expressing our gratitude to Professor LeCun for his invaluable contributions to the field of AI and for his inspiring leadership! ✨
#AI #GenerativeAI #WinterSchool #EminesUM6P #YannLeCun #TuringAward #Meta #NYU #MachineLearning #Education #Innovation - Наука та технологія
David Deutsch also composes the energy equation as split by work medium plus heat medium. Where heat medium is the entropy value or complexity you dont understand that is wasted. Work medium is structure that you know and can exploit.
Possibly the brain orchestrates hierarchical planning through superpositions of frequency modulation where slow frequency channels are abstract high frequncy channels are high level detail. We need similar orchestrated messaging for ad hoc contexts.
Energy models like constructor theory where high value scalars reflect level of impossibility. Low scalars reflect proximity to possibility.
Conjecture could include deduction and abduction, which i believe is what is wanted here.
Can we call inference conjecture. The word inference includes induction, abduction and deduction.
Can training max entropy trade off be generalized to just trade off? For example arent guard rails trade offs. Isnt all control just trade off of a limited resource?
Does the brain run everything in reverse during sleep to represent locally impossible to prevent collapse?
Is it just me or does the sound start out fine and get really bad a few seconds in?
It's happening for me as well. Difficult to hear clearly when on max volume
Hello, strange at my place everything is working fine... I uploaded a version with enhanced sound via this link: ua-cam.com/video/58881_mGm94/v-deo.htmlfeature=shared. Enjoy watching and sorry!
@@melrhabi Thank you, that's much better.
not audible, unfortunately :(
Hello, strange at my place everything is working fine... I uploaded a version with enhanced sound via this link: ua-cam.com/video/58881_mGm94/v-deo.htmlfeature=shared. Enjoy watching and sorry!"
Brain limited to one planner and only one geometry navigation model in hippocampus
Machines don’t surpass human intelligence but may surpass performance in certain tests. In any case these tests are misguided the same you don’t test a car using human benchmarks.