Variational Dropout – Uncertainty in Deep Learning?

Note: This article was originally published on September 8, 2017 and has been migrated from our previous blog. Some details — tools, libraries, benchmarks, industry context — may be outdated. For our latest perspective, see our recent posts.

Last Thursday, our Senior Deep Learning Researcher gave a lecture in Warsaw at the event co-sponsored by Sigmoidal – Data Science PL group:

Marcin covers the topic of measuring certainty using a relatively unknown technique known as Variational Dropout.

Another method only briefly covered in research that proved very useful in working with small datasets coming from our clients is called Positive Learning.

Watch the full video

This video begins a series of #meetsigmo video lectures, where we aim to raise awareness about less known topics in the Deep Learning ecosystem that proved valuable for us.

Be sure to subscribe to our YouTube channel to stay up-to-date with the latest research.