Chapter
The Challenge of Context in Deep Learning
32:40 - 37:24 (04:44)
The contextual comprehension is what makes tasks hard for deep learning systems. The challenge lies exactly in the things they were good at, such as image and speech recognition, while the human brain excels at solving seemingly complex problems like chess, due to its contextual understanding.
Clips
This podcast episode discusses the importance of symbol grounding and the relationship between objects in deep learning systems to achieve color constancy, using examples such as the misclassification of a stop sign with a few marks on it.
32:40 - 34:10 (01:29)
Summary
This podcast episode discusses the importance of symbol grounding and the relationship between objects in deep learning systems to achieve color constancy, using examples such as the misclassification of a stop sign with a few marks on it.
ChapterThe Challenge of Context in Deep Learning
Episode#217 – Rodney Brooks: Robotics
PodcastLex Fridman Podcast
This podcast episode features a conversation about the challenges of perception and reasoning in AI, and how humans have a poor intuition of what is hard and what is easy.
34:10 - 35:36 (01:26)
Summary
This podcast episode features a conversation about the challenges of perception and reasoning in AI, and how humans have a poor intuition of what is hard and what is easy.
ChapterThe Challenge of Context in Deep Learning
Episode#217 – Rodney Brooks: Robotics
PodcastLex Fridman Podcast
The notion that hard things necessarily equate to the highest intellect is challenged with examples from chess and the rise of artificial intelligence.
35:36 - 37:24 (01:48)
Summary
The notion that hard things necessarily equate to the highest intellect is challenged with examples from chess and the rise of artificial intelligence.