Learning Computational Structure for Sample-Efficient Prediction
The authors study whether gradient-based prediction can exploit hidden compositional (graph) structure. They introduce CSFRBench, which hides the target computation during training but keeps it for diagnosis, and propose Learned Wiring (LeWi): learn low-dimensional neuron embeddings that determine which connections are active (within a radius) while jointly training connection weights. LeWi improves sample efficiency on hierarchical, reuse, or modular targets and performs better than a dense baseline on 17 world-model prediction tasks.
