concept
active
concept:learnable-via-data-scaling-definition-1Learnable via data scaling (Definition 1)
Loss gains a smaller model recovers with infinite data.
Neighborhood — ranked by edge-count
Papers (1)
paper
Concepts (1)
concept
- Learnable via model scaling (Definition 2)associated_withrelated_toLoss a smaller model never reaches even asymptotically.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Interpretive process for transforming many-valued contexts into formal contexts via scale attributes.
- Used in the color cooccurrence experiment to embed colors into 3D space preserving dissimilarity matrix distances
- Process by which low-level competencies scale to larger problem-solving abilities across levels of organization.
- Inference of parameters encoding contingencies of the world (e.g., likelihood matrix A) at slower timescale than perception.
- Clarifies what unsupervised learning does.
- Field of research integrating reward learning and optimization; shown to be unified with perceptual learning via free energy principle.
- Key insight about predictive learning's potential.
- Authors argue features are model properties because logit effects and ablations are consistent with feature interpretations