Meta-Learning View of ICL (Dai et al. 2023)

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Optimization-as-Inference View of ICL (von Oswald et al. 2023)framework0.826
  Views transformers as performing implicit gradient descent; UCCT complements this mechanistic account
• in-context learning (ICL)concept0.791
  Test-time adaptation from prompt or retrieved context with no parameter updates.
• Bayesian ICL (Xie et al. 2022)framework0.785
  Prior framework explaining ICL as inference over task structure; UCCT adopts and extends the Bayesian lens
• Meta-learningconcept0.772
  The capability of GPT-3 to learn tasks from few-shot prompts during runtime.
• Causal Representation Learning (CRL)framework0.737
  Schölkopf et al.'s framework combining representation learning with causal inference.
• Competition Dynamics ICL Framework (Park et al. 2024)framework0.727
  Formalizes regime shifts between retrieval-like and inference-like ICL; UCCT complements with when-predictor
• Critical social learning—where humans evaluate and override AI—is the only strategy recovering collective performance.claim0.723
• If successful, CRL could enable AI systems to not only replicate human contemplative practices but also generate novel potentially superhuman forms of contemplative and ethical reasoningclaim0.721
  Ambitious claim comparing CRL potential to AlphaGo's move 37 in game-playing