Unverified70% confidenceFactExact time
In "On Calibration of Modern Neural Networks" published at ICML in 2017 by Guo et al., the miscalibration phenomenon was empirically demonstrated at scale for the first time, finding that modern deep networks such as ResNet are worse calibrated than earlier shallow networks.
2
Sources
70%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
元认知反馈：用强化学习让大模型准确表达不确定性
hackernewshackernews7/7/2026
Related Claims
UnverifiedResidual connections were first introduced into deep learning by He Kaiming in ResNet in 2015.74% similarVerifiedResNet (Residual Network) was proposed by Microsoft Research's He Kaiming team in 2015 and solved the vanishing gradient problem in deep network training by introducing residual connections.73% similarUnverifiedTreating UAT as proof that "deep networks are superior to shallow networks" is a common misconception.69% similarUnverifiedResNet（残差网络）通过跳跃连接解决了深层网络的梯度消失问题，能提取层次化视觉特征68% similarUnverifiedExperiments by Zhang et al. (2017) demonstrated that standard deep neural networks can perfectly memorize training data with random labels, meaning that traditional generalization bounds based on model capacity completely break down.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/279890API
curl https://kongchang.com/api/v1/knowledge/claims/279890MCP
get_claim(id=279890)