In reply to UVM_SV_101:
Machine learning(ML) has become a big buzzword in the past year, just like Internet of things(IoT) was a few years ago, and dot com before that. Back then you could get a 10 fold increase in funding just by adding .com to your startup’s name. Right now, people are trying to apply ML to almost anything. That said I can think of two areas that might be specific to Design Verification.
Companies with huge data warehouses have been active in developing hardware aiding ML algorithms for a while. Design knowledge always helps with verification.
There is also activity in applying ML algorithms to aid functional coverage closure. There is massive amounts of data generated from regressions from massive amounts of tests. I’ve seen papers at DVCon from people trying to guide constrained random stimulus, triage regression failures, and shorten regression cycles. Unfortunately, the proceedings are offline while there is a transition in conference management.