Search

Tag: #contrastive-learning6 results

CWM提升具身代理動作可行性

CWM提升具身代理動作可行性

CWM使用InfoNCE對比學習與硬挖掘負例,對LLM進行微調作為動作評分器,以區分可行與無效動作。它在最小編輯硬負例上Precision@1優於SFT達+6.76個百分點,並達成更高AUC-ROC(0.929)。在ScienceWorld即時測試中,CWM在OOD壓力下提供更好安全邊際。

FASCL Future-Aligns Asset Retrieval

FASCL Future-Aligns Asset Retrieval

FASCL employs future-aligned soft contrastive learning using pairwise return correlations as supervision for financial asset retrieval. It outperforms historical similarity baselines on US equities. Includes protocol to evaluate future trajectory alignment.

ArXiv AIResearchFeb 12#research#fascl#v1