[KongchangAI]
Concept

Cross-Group Reconstructability

A selective constraint mechanism in collaborative learning that uses disjoint groups to reconstruct task-response increments, forcing a model to extract useful information from a collaborator's representation rather than directly copying its outputs.

Timeline (last 90 days)

Nov 2

In CL, each model generates task responses and reconstructs response increments using collaborators' representations, but the reconstruction process excludes source information, separating fitting and evaluation on disjoint groups.

Unverified50%

All Facts (1)

Source Articles