RAL models PRVR with probabilistic Gaussian alignment plus confidence-weighted word matching, improving SumR by 9.7 over prior best on TVR.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Partially Relevant Video Retrieval with Robust Alignment Learning
RAL models PRVR with probabilistic Gaussian alignment plus confidence-weighted word matching, improving SumR by 9.7 over prior best on TVR.