AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Justification-Based Reliability in Machine Learning

Nurali Virani, Naresh Iyer, Zhaoyuan Yang

AAAI 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use ε-ball support rather than softmax thresholdingGrid Stability FC epistemic classifier · Grid Stability, large-noiseA_IK: 0.433 -> 0.750Improved
Use ε-ball support rather than softmax thresholdingItaly Power Demand CNN epistemic classifier · Italy Power Demand, large-noiseA_IK: 0.712 -> 1.000Improved
Use ε-ball support rather than softmax thresholdingMNIST FC epistemic classifier · MNIST, adversarial noiseF_IK: 0.851 -> 0.000Improved
Use ε-ball support rather than softmax thresholdingGTSRB CNN epistemic classifier · GTSRB, adversarial noiseF_IK: 0.631 -> 0.140Improved
Use ε-ball support rather than k-NN supportSynCon CNN epistemic classifier · SynCon, adversarial noiseF_IK: 0.440 -> 0.343Mixed
Use ε-ball support rather than k-NN supportMNIST FC epistemic classifier · MNIST, adversarial noiseF_IK: 0.222 -> 0.000Improved
Use ε-ball support union k-NN support (H-2) instead of ε-ball support (ablation)SynCon CNN epistemic classifier · SynCon, adversarial noiseA_IK: 0.942 -> 0.957Improved
Use ε-ball support union k-NN support (H-2) instead of ε-ball support (ablation)GTSRB CNN epistemic classifier · GTSRB, adversarial noiseF_IK: 0.140 -> 0.102Hurt
Use H-1 fallback to k-NN when the ε-ball is empty instead of ε-ball support (ablation)SynCon CNN epistemic classifier · SynCon, adversarial noiseA_IK: 0.942 -> 0.942No clear effect
Use support from a logit layer only instead of an intermediate layer plus logit (ablation)SynCon CNN epistemic classifier · SynCon, adversarial noiseA_IK: 0.957 -> 0.699Hurt
Use support from a logit layer only instead of an intermediate layer plus logit (ablation)MNIST FC epistemic classifier · MNIST, adversarial noiseA_IK: 0.000 -> 0.000No clear effect
Use support from a logit layer only instead of an intermediate layer plus logit (ablation)GTSRB CNN epistemic classifier · GTSRB, adversarial noiseA_IK: 0.018 -> 0.028Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy