Cooperation, Competition and RSI

Presentation Link#

The seventy-forth MeetUp of the Machine Learning Singapore Group, was titled : “MLSG - Reinforcement Learning : DIY and RSI”, and we covered some of the latest developments in making models better along the road to AGI with new advances in model training, along with an introductory talk about training LLMs using Reinforcement Learning (RL) on your own data.

My Presentation#

My talk was titled “Competition of Cooperation” - and was motivated by the increasing number of interesting combinations of training an LLM by employing another LLM in a different role. Many of these techniques can be traced back to pre-ChatGPT work (for instance: AlphaZero or DINO), so I went over those methods as a warm-up.

Overall, I used the following outline :

• What’s the game?
• Some history
  • pre-LLM
• Modern day
  • exciting directions
• Other News…
• Wrap-up & QR-code

Slides#

The slides for my talk, which contain links to all of the reference materials and sources, are here :

If there are any questions about the presentation please ask below, or contact me using the details given on the slides themselves.

Other Presentations#

The MeetUp also featured Sam Witteveen giving an introductory (“Part 1”) talk about training your own (local) LLMs via Reinforcement Learning.

Acknowledgements#

Many thanks to the Google team, who allowed us to use Google’s Developer Space, and generously provided food for our audience.