Will a hierarchical RL algorithm be SOTA on any major RL benchmark by 2025?
3
13
Ṁ32Ṁ160
2025
38%
chance
1D
1W
1M
ALL
I'm intentionally being somewhat vague. Certainly if someone released an option-based algorithm that achieved SOTA on Atari that would count. On the flip side an algorithm that achieved SOTA on a benchmark specifically designed to test hierarchical learning wouldn't count.
Get Ṁ200 play money
Related questions
What will be true of OpenAI's Sora* model, at the end of 2025? [*see description]
SoAI 23 3/10: Will Self-improving Al agents crush SOTA in a complex environment (e.g. AAA game, tool use, science)?
29% chance
Will any AI be able to formalize >=90% of IMO problems by the start of 2025?
32% chance
Will openAI have the most accurate LLM across most benchmarks by EOY 2024?
39% chance
Will Transformer based architectures still be SOTA for language modelling by 2026?
67% chance
Will reinforcement learning overtake LMs on math before 2028?
43% chance
Will OpenAI, Google, or Anthropic Have a SOTA Model's Weights Hacked Before 2025?
26% chance
Will >50% of the tasks in the WebArena benchmark be solved by EOY 2024?
62% chance
Will an AI agent system be able to score at least 40% on level 3 tasks in the GAIA benchmark before 2025.
49% chance
Will self-improving AI agents crush SOTA in a complex environment (e.g. AAA game, tool use, science) in next 12 months?
41% chance