Metadata#
- Author(s): Patrick Henry Winston
- Number of pages: 527
- Year published: 1984
- Year read: 2018
Review#
DNF at about halfway, and won’t review, since - meh - it was fine. I used it during my first Georgia Tech OMSCS course, Knowledge-Based Artificial Intelligence, and that’s over now, so… 🤷
I mean, I love Winston’s MIT AI lectures. They are MIND BLOWINGLY good. Here they are. This book is less about “optimal” AI (in the style of neural nets and machine learning) and more about “human-like” AI (in the style of mimicking cognitive systems and encoding semantic relationships). Which I find inherently confusing and kind of gratingly dull. ML is more exciting! And seems to make more “sense” (even if it’s less interpretable, often). As such, this book feels like a textbook from history - from a field that has changed quite a lot since it was published.