Professor Yann LeCun recently gave a talk titled "From Machine Learning to Autonomous Intelligence"
The presentation covers an important topic:
How could machines learn, remember, reason, plan, and have common sense, yet be steerable and safe?
Here is a summary:
1. "AI will affect every aspect of human activity"
There are numerous benefits of AI in many different areas of science and everyday life.
From medicine, transportation to arts and education.
2. "Self-Supervised Learning has taken over the world"
Models like BERT and other foundation models have changed the way we use machine learning.
They can be used for "understanding and generating text, images, video, 3D models, speech, proteins, and more."
3. Auto-Regressive Large Language Models suck
Performance is amazing but
- They make stupid mistakes
- They have limited knowledge of the underlying reality
- They have no common sense
- They can’t plan their answer
4. Current AI Technology is (still) far from Human Level
"There is no questions that, eventually, machines will eventually surpass human intelligence in all domains"
5. What's more in the talk:
- Three challenges for AI & Machine Learning
- Objective-Driven AI
- How could machines learn to understand the world?
- Energy-Based Models
- What is AGI? How long is it going to take to reach human-level AI?
- Will robots take over the world?
All the information is taken from the talk recording and slides.
Video: https://lnkd.in/evj6nxtb
Slides: https://lnkd.in/eXpAXHMf
Financial Services Professional
1wAn innovative solution for preventing Fog Accidents using AI. https://youtu.be/7CyuBkBTeAg?si=0S_WIdMWnQpHT058