🤖 MLguru #13: Fighting DeepFakes, Lung Cancer Detection and Reinforcement Learning in Games

Photo of Mateusz Opala

Mateusz Opala

Updated Jan 4, 2023 • 4 min read
MLguru - quotes (1)

Two weeks have passed and it’s time for the new, 13th edition of MLguru, our bi-weekly Machine Learning newsletter. In this edition, you will read about:

  • Facebook and it’s input in fighting DeepFakes,
  • Lung cancer detection,
  • Framework for general reinforcement learning with 20+ games,
  • The state of transfer learning in NLP,
  • And much more!

If you have any suggestions for future editions, drop me a line. I’ll be more than happy to get your feedback! In return, I’ll share the recording of our recent ML webinar: How to build and implement Machine Learning Models.

DeepFakes Detection Challenge

DeepFakes are unfortunately becoming more and more popular, and their use might be very harmful. This is why Facebook is partnering with academic researchers and other industry leaders to create the DeepFake Detection Challenge.

The goal of the challenge is to build a technology that everyone could use to successfully detect if AI has been used to alter a video in order to mislead the viewer. The Deepfake Detection Challenge will include a data set and a leaderboard, but it will also offer grants and awards to motivate developers to create new ways of detecting and preventing media manipulated via AI from being used to mislead others. Read more.

ML contributes to lung cancer detection

Google AI researchers have recently created an AI model capable of detecting lung cancer from screening tests better than experienced radiologists. The machine learning model detected cancer 5% more often on average than a group of six human experts and was 11% more likely to reduce false positives.

If you are interested in healthcare technological solutions, read our recent report to learn about patient management software market.

OpenSpiel: Framework for general reinforcement learning with 20+ games

OpenSpiel is a collection of algorithms and environments for research in general reinforcement learning and search/planning in games. OpenSpiel supports n-player (single- and multi- agent) zero-sum, cooperative and general-sum, one-shot and sequential, strictly turn-taking and simultaneous-move, perfect and imperfect information games, as well as traditional multiagent environments such as (partially- and fully- observable) grid worlds and social dilemmas. You can read more about the project here.

The State of Transfer Learning in NLP

Sebastian Ruder, Matthew Peters, Swabha Swayamdipta and Thomas Wolf have created a great tutorial showing the state of transfer learning in NLP. It’s a complex post for everyone willing to learn or broaden the knowledge on Transfer Learning and Natural Language Processing. It’s available in both English and Chinese, and you can find it here.

DistilBERT, a distilled version of BERT

In our previous editions of MLguru, we told you about BERT and it’s younger sister roBERTa. Now it’s time for DistilBERT, which is smaller, faster, cheaper, and lighter.

MLguru - quotes (1)

Read more about distilBERT on Medium.

Photo of Mateusz Opala

More posts by this author

Mateusz Opala

How to build products fast?  We've just answered the question in our Digital Acceleration Editorial  Sign up to get access

Read more on our Blog

Check out the knowledge base collected and distilled by experienced professionals.

We're Netguru!

At Netguru we specialize in designing, building, shipping and scaling beautiful, usable products with blazing-fast efficiency
Let's talk business!

Trusted by: