Per-feature coordinate-search thresholds for binarizing BERT embeddings outperform fixed-threshold methods and occasionally match full-precision accuracy.
Google Landmark Recognition 2020 Competition Third Place Solution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We present our third place solution to the Google Landmark Recognition 2020 competition. It is an ensemble of global features only Sub-center ArcFace models. We introduce dynamic margins for ArcFace loss, a family of tune-able margin functions of class size, designed to deal with the extreme imbalance in GLDv2 dataset. Progressive finetuning and careful postprocessing are also key to the solution. Our two submissions scored 0.6344 and 0.6289 on private leaderboard, both ranking third place out of 736 teams.
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Evolutionary Feature-wise Thresholding for Binary Representation of NLP Embeddings
Per-feature coordinate-search thresholds for binarizing BERT embeddings outperform fixed-threshold methods and occasionally match full-precision accuracy.