Unverified50% confidenceOpinionExact time
Zero-shot transfer may still be limited by differences between target datasets in scene types, annotation density and motion patterns, and practical deployment often still requires a small amount of adaptation.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
Unverified当目标数据域与源数据域差异悬殊时，预训练权重可能产生负迁移（Negative Transfer），反而干扰目标域的学习69% similarUnverifiedIn transfer learning practice, the smaller the amount of data for the new task, the more layers should be frozen to prevent overfitting; the learning rate should be much smaller than when training from scratch; if the new data differs greatly in distribution from the pretraining data, more layers often need to be fine-tuned.68% similarUnverifiedTRACER achieves the lowest values on both group-level conversion rate error and semantic trajectory distance metrics, and generalizes to out-of-distribution scenarios.68% similarUnverified应对域偏移的常见策略包括少量标注数据的微调、无监督域适应以及测试时增强68% similarUnverifiedCovariate shift and model drift can trigger negative transfer, causing additional data to drag down model performance.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1041858API
curl https://kongchang.com/api/v1/knowledge/claims/1041858MCP
get_claim(id=1041858)