MLSG April Meetup - Bringing You News from the Frontier

This month's event is just after Google Cloud Next (in Las Vegas) and ICLR (academic conference in Singapore). Talks may be subject to change!

Talks:

"Latent Space Reasoning" - Martin Andrews

Reasoning using LLMs has taken off since the release of Deepseek's R1 model in January. But are tokens the best way of expressing a line of thought? Martin will briefly outline the various research directions for this new direction - and also add some of his ICLR take-aways...

"Open-endedness and Interestingness" - Jenny Zhang

At the standing-room-only World Models workshop at ICLR, Jeff Clune gave an inspiring talk about Open-endedness and the path beyond standard benchmarks. Several of the key papers he showed had Jenny Zhang as a primary author - and we are proud to welcome her to talk about her ICLR takeaways while she's back in Singapore.

"My ICLR Highlights" - Raymond Chan

Returning to speak at MLSG, Raymond will be going through 3 papers that he found interesting from the Orals/Spotlights from ICLR. The paper's titles include the words : "Trolleys", "Irrationality" and "Sensitive Images".

"Agent Announcements & Trends from Google Cloud Next 2025 " - Sam Witteveen

In this talk, Sam will cover some of the key announcements from Google's Cloud Next 25 event in Las Vegas. These include the new Agent Development Kit (ADK), Agent Space and also the new Agent2Agent protocol. Sam will also talk about some of the trends seen from startups and companies presenting their agentic products there.

---
Head's up! The "hands-on" Google workshop on the afternoon of Saturday 3-May : "Advanced Build with AI" has had to be POSTPONED due to Singapore's recently announced General Election! We'll update everyone once we know the new date for the workshop...
---

Talks will start at 7:00pm and end at around 8:45pm, at which point people normally come up to the front for a bit of a chat with each other, and the speakers.

As always, we're actively looking for more speakers - both '30 minutes long-form', and lightning talks. For the lightning talks, we welcome folks to come and talk about something cool they've done with keras, PyTorch, JAX and/or Deep Learning for 5-10mins (so, if you have slides, then #max=10). We believe that the key ingredient for the success of a Lightning Talk is simply the cool/interesting factor. It doesn't matter whether you're an expert or an enthusiastic beginner: Given the responses we have had to previous talks, we're sure there are lots of people who would love to hear what you've been playing with. If you're interested in talking, please just introduce yourself to Martin at one of the events.