TRACE achieves 97% macro F1 on temporal out-of-distribution prediction of organizational exploit targets using contrastive learning on a 129k-sample multi-source dataset, outperforming 17 baselines.
Detecting Cyber Threats in Non-English Dark Net Markets: A Cross-Lingual Transfer Learning Approach,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2020 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Vendor-Conditioned Contrastive Learning for Predicting Organizational Cyber Threat Targets
TRACE achieves 97% macro F1 on temporal out-of-distribution prediction of organizational exploit targets using contrastive learning on a 129k-sample multi-source dataset, outperforming 17 baselines.