The paper reformulates situation-recognition verb classification as single-positive multi-label learning and reports a 25,200-image multi-label benchmark with a graph and adversarial training method that improves MAP by about 3 points over the multi-class baseline.
Gandef: A gan based adversarial training defense for neural network classifier,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
The Demon is in Ambiguity: Revisiting Situation Recognition with Single Positive Multi-Label Learning
The paper reformulates situation-recognition verb classification as single-positive multi-label learning and reports a 25,200-image multi-label benchmark with a graph and adversarial training method that improves MAP by about 3 points over the multi-class baseline.