A capsule network with dynamic routing is reported to classify SNIPS intents with 95.6% accuracy, but uncontrolled baselines and missing experimental details undermine the claim.
A multiscale CNN-CRF framework for environmental microorganism image segmentation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Capsule Network-Based Semantic Intent Modeling for Human-Computer Interaction
A capsule network with dynamic routing is reported to classify SNIPS intents with 95.6% accuracy, but uncontrolled baselines and missing experimental details undermine the claim.