MinedDeep

David Silver

2 min read

David Silver

Type: collaborator Slug: collaborator—david-silver Sources: human-level-control-through-deep-reinforcement-learning—hassabis, mastering-the-game-of-go-with-deep-neural-networks-and-tree-search—hassabis, mastering-the-game-of-go-without-human-knowledge---hassabis, a-general-reinforcement-learning-algorithm-that-masters-chess-shogi-and-go—hassabis, grandmaster-level-in-starcraft-ii-using-multi-agent-rl—hassabis, mastering-atari-go-chess-and-shogi-by-planning-with-a-learned-model—hassabis, highly-accurate-protein-structure-prediction-with---hassabis, reinforcement-learning-fast-and-slow—hassabis Last updated: 2026-05-13

Summary

David Silver is the most frequent co-author in the DeepMind era, appearing on 8 corpus papers (2015–2021) spanning the entire game-playing arc (DQN → AlphaGo → AlphaGo Zero → AlphaZero → MuZero → AlphaStar) and the AlphaFold2 paper. He is the intellectual lead of DeepMind’s reinforcement learning programme.

Role

Lead researcher on RL at DeepMind. First author on no corpus paper but consistent co-author across the entire game-playing sequence. His UCL PhD thesis (on reinforcement learning with tree search) provided the theoretical foundation for AlphaGo’s MCTS integration.

Papers together

  1. Human-level control through deep RL (2015, DQN)
  2. Mastering the game of Go (2016, AlphaGo)
  3. Mastering the game of Go without human knowledge (2017, AlphaGo Zero)
  4. A general RL algorithm (2018, AlphaZero)
  5. Grandmaster level in StarCraft II (2019, AlphaStar)
  6. Mastering Atari, Go, chess and shogi (2020, MuZero)
  7. Reinforcement learning, fast and slow (2021)
  8. Highly accurate protein structure prediction (2021, AlphaFold2)

Connections

  • Projects: project—DQN, project—AlphaGo, project—AlphaFold2
  • Themes: theme—deep-RL, theme—self-play, theme—game-playing-AI
  • Periods: period—early-deepmind through period—alphafold-era

Honest Gaps

  • No interview or first-person account from Silver in the corpus — his intellectual contributions are inferred from authorship positions.
  • The Silver-Hassabis division of labour is never described.
  • Silver’s own prolific publication record (beyond co-authorship with Hassabis) is not represented.

Graph View

Backlinks