Hacker News new | past | comments | ask | show | jobs | submit login
Steps Toward Robust Artificial Intelligence [pdf] (oregonstate.edu)
16 points by fitzwatermellow on Feb 16, 2016 | hide | past | favorite | 2 comments

Nice viewgraphs! I hope I can find the video of the talk on the ACM web site (I couldn't find it on youtube).

With all of the press about deep learning it is also good to see progress in other AI fields (1 million constraints now for SAT solvers, etc.)

I saved the PDF in my permanent archive - really good stuff.

Class Detection via Anomaly Detection shows really impressive performance.

Anybody got an intuition as to why is that so? Something like having opinionated low-variance high-bias classifiers that you ensemble? In which case, how can it be better than random forests?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact