S3 decomposes multimodal data into selectable semantic experts, routes them adaptively, and sparsifies to achieve higher accuracy on MultiBench benchmarks with peak performance at intermediate sparsity levels.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Derives information-maximizing rules for baseline weights and release probabilities in Tsodyks-Markram synapses, producing onset-sensitive presynaptic terms and anti-causal connectivity in recurrent networks.
citing papers explorer
-
Toward Structural Multimodal Representations: Specialization, Selection, and Sparsification via Mixture-of-Experts
S3 decomposes multimodal data into selectable semantic experts, routes them adaptively, and sparsifies to achieve higher accuracy on MultiBench benchmarks with peak performance at intermediate sparsity levels.
-
Reshaping Neural Representation via Associative, Presynaptic Short-Term Plasticity
Derives information-maximizing rules for baseline weights and release probabilities in Tsodyks-Markram synapses, producing onset-sensitive presynaptic terms and anti-causal connectivity in recurrent networks.