Location: Master Stage
DeepSeek R1 signals a significant turning point of LLM research and development, where more new architecture improvements utilised in pre-training and a brand new RL flywheel in post-training that desires the power of open source.
In this talk we will first introduce LF AI & Data Foundation, then walk through six directions and trends of LLM research and finally we present a proposal of building open source ecosystem for reinforcement learning with generalised reward in a new era of open experience
In this talk we will first introduce LF AI & Data Foundation, then walk through six directions and trends of LLM research and finally we present a proposal of building open source ecosystem for reinforcement learning with generalised reward in a new era of open experience
Speakers
Share:
