NeurIPS Roundup

Presentation Link

The sixty-seventh MeetUp of the Machine Learning Singapore Group, was titled : "NeurIPS Recap - The State of AI Research" - and had talks covering the latest in AI research.

My Presentation

My talk was titled "Nested Learning, mHC / Engram, & Sub-1bit Compression" (a grab-bag of interesting research from NeurIPS in Dec-2025 and the month following). It had the following outline :

• Intro
• Three interesting directions:
  • "mHC" and "Engram" - DeepSeek
  • "Nested Learning" - Google
  • "LittleBit: Ultra Low-Bit Quantization" - Samsung
• Heads-Up!
  • Sakana "DroPE"
  • Google: "Prompt Repetition Improves Non-Reasoning LLMs"
• Wrap-up & QR-code (the latter to reduce audience distractions)

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 included two other speakers: Rach Pradhan and Sam Witteveen:

• In his lightning talk, Rach described his highly-rated entry into the recent DeepMind Gemini Hackathon held in Singapore. His Open Source "EmergentDB" is a proof-of-concept vector database that uses MAP-Elites (Quality Diversity) to automatically discover optimal index configurations

• Sam went through some of the papers and topics covered at NeurIPS and in general research over the past few months that cover how in the race for AGI, research is moving to Agentic Systems rather than just Big Models

Acknowledgements

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