Unverified50% confidenceFactExact time
Koh & Liang introduced influence functions into deep learning in 2017, using them to identify mislabeled samples and understand model behavior.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedKernel ridge regression (KRR) is an algorithm with a solid theoretical foundation in nonparametric learning and is often used to study the intrinsic properties of deep learning.72% similarUnverifiedDeep CFR uses a function approximator (neural network) to directly learn the mapping from information sets to strategies/regret values, and was proposed by Brown et al. in 2019.69% similarUnverifiedShortcut learning is a widely observed phenomenon in deep learning, where models use features that are statistically correlated with labels but semantically irrelevant to make predictions.68% similarUnverifiedGrinsztajn等人2022年在NeurIPS发表的基准研究证明，在中等规模表格数据集上，GBDT类模型持续优于深度学习方法67% similarUnverified用极小数据子集验证模型能否过拟合是调试深度学习模型的第一步，由 Andrej Karpathy 在《A Recipe for Training Neural Networks》中系统阐述65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1022903API
curl https://kongchang.com/api/v1/knowledge/claims/1022903MCP
get_claim(id=1022903)