Reproduce “Connecting Context-specific Adaptation in Humans to Meta-learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Rachit Dubey, Erin Grant, Michael Luo, Karthik Narasimhan, Thomas Griffiths
MLTI lets task context shape a learner’s initialization before task-specific examples or rewards drive adaptation. A context network produces a conditioned starting point, while online feedback still updates the learner. In an infant-learning task reproduction, its validation loss was 3.65 and 6.19 in consistent and inconsistent contexts, unlike MAML; simulated benchmark gains were task-dependent, not universal.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.