question
active
question:how-can-subcomponents-be-algorithmically-clustered-into-full-components-when-ground-truth-is-unknownHow can subcomponents be algorithmically clustered into full components when ground truth is unknown?
Open problem.
Source paper
extracted_from(2025) · Lucius Bushnaq · Dan Braun · Lee Sharkey
Neighborhood — ranked by edge-count
Papers (1)
paper
- Stochastic Parameter Decompositionassociated_withmentions
Claims (1)
claim
- Central claim.
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.
- Tested in Section 4.4 calibration experiment; confirmed by findings.
- Central claim from connectionist models: complex coordination emerges without centralized control or external teacher.
- Layer choice discussion.
- Motivated by the finding that lexical entailment decomposes into word identities.
- Assertion about the qualitative advantages of VPD's rank-one decomposition.
- Discovery story.
- Interpretive claim that the subcomponents correspond to real functional units.
- Future work.