Chapter
Advancing Artificial Intelligence Research beyond Passively Observing Data
07:48 - 20:28 (12:39)
The advancement of artificial intelligence research into active agents that learn by intervening in the world and the development of objective functions that allow for high-level explanations to emerge from the learning process present new and exciting opportunities for academics to advance the state of the art, particularly in training frameworks, learning models, and agent learning in synthetic environments, using projection of data into the right semantic space.
Clips
The speaker highlights the opportunities for academics to contribute to AI research in terms of advancing learning models, agent learning, and training frameworks, especially for synthetic environments where even state-of-the-art deep learning methods struggle to understand even simple models.
07:48 - 15:22 (07:33)
Summary
The speaker highlights the opportunities for academics to contribute to AI research in terms of advancing learning models, agent learning, and training frameworks, especially for synthetic environments where even state-of-the-art deep learning methods struggle to understand even simple models.
ChapterAdvancing Artificial Intelligence Research beyond Passively Observing Data
EpisodeYoshua Bengio: Deep Learning
PodcastLex Fridman Podcast
By projecting data into a semantic space, it is possible to represent extra knowledge beyond the transformation from inputs to representations in a way that can be neatly disentangled.
15:22 - 20:28 (05:06)
Summary
By projecting data into a semantic space, it is possible to represent extra knowledge beyond the transformation from inputs to representations in a way that can be neatly disentangled. The author argues that learning about the relationships between variables in high-level semantic variables is an important next step in improving current neural nets.