Obstacles to progress in AI
The essence of intelligence is the ability to predict. Prediction, perception, planning/reasoning, attention, and memory are the pillars of intelligence. Yann LeCun describes several projects at FAIR and NYU on unsupervised learning, question answering with a new type of memory-augmented network, and various applications for vision and natural language understanding.
Talk Title | Obstacles to progress in AI |
Speakers | Yann LeCun (Facebook) |
Conference | O’Reilly Artificial Intelligence Conference |
Conf Tag | |
Location | New York, New York |
Date | September 26-27, 2016 |
URL | Talk Page |
Slides | Talk Slides |
Video | |
The essence of intelligence is the ability to predict. Prediction, perception, planning/reasoning, attention, and memory are the pillars of intelligence. Both animals and humans learn to predict, learn how the world works, and acquire common sense largely without supervision, through observation and experimentation. This is a far cry from supervised learning—the basis of most recent successes in the application of deep learning. Significant progress in AI will require breakthroughs in unsupervised/predictive learning, as well as in reasoning, attention, and episodic memory. Yann LeCun describes several projects at FAIR and NYU on unsupervised learning for predicting videos using adversarial training, question answering with a new type of memory-augmented network, and various applications for vision and natural language understanding.