Research map / Adversarial robustness and security / Attacks on text and graphs
Enhancing Multiple Reliability Measures via Nuisance-Extended Information Bottleneck
Jongheon Jeong, Sihyun Yu, Hankook Lee, Jinwoo Shin
This paper tried a common change: Add a reconstruction / auxiliary loss
Add the reconstruction loss · CIFAR-10 · accuracy Improved
What 987 accepted papers report for this change
- 882 improved
- 41 no clear effect
- 20 mixed
- 44 hurt
- OPTION: Optimal Transport–Guided Flow Matching for Incomplete and Unaligned Multi-View Clustering ICML 2026 · add reconstruction loss L_rec · Scene-15 (aligned, 70% missing) · ACC: 32.4 -> 37.5Improved
- Detached Skip-Links and $R$-Probe: Decoupling Feature Aggregation from Gradient Propagation for MLLM OCR ICML 2026 · add R-probe auxiliary loss · OCRBench · score: 714.0 -> 721.0Improved
- Rethinking Time-Series Imputation as Conditional Inference along Temporal Evolution ICML 2026 · add reconstruction loss on observed entries (joint vs missing-only supervision) · ETTh1 10% missing · Masked MAE: 0.2002 -> 0.1658Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Train with AENIB rather than cross-entropy | ResNet-18 · CIFAR-10-C | average corruption error rate: 16.0 -> 12.3 | Improved |
| Train with AENIB rather than PixMix | ResNet-18 · CIFAR-10.1 | test error rate: 13.1 -> 11.6 | Improved |
| Train with AENIB rather than the cross-entropy baseline | ViT-S/16 · ImageNet-R | error rate: 70.3 -> 67.1 | Improved |
| Train with AENIB rather than the cross-entropy baseline | ViT-S/16 · Background Challenge Only-BG-T | error rate: 20.3 -> 17.8 | Improved |
| Add nuisance likelihood to the Dirichlet novelty score (ablation) | AENIB · OBJECTS MNIST | AUROC: 90.53 -> 92.43 | Improved |
| Add nuisance likelihood to the Dirichlet novelty score (ablation) | AENIB · OBJECTS Texture | AUROC: 75.04 -> 88.91 | Improved |
| Replace NMSE reconstruction loss with D2^2 (ablation) | AENIB ViT-S/4 · CIFAR-10-C | average corruption error rate: 13.0 -> 12.3 | Improved |
| Replace NMSE reconstruction loss with D2^2 (ablation) | AENIB ViT-S/4 · CIFAR-100-C | average corruption error rate: 36.9 -> 35.8 | Improved |
| Increase the information-bottleneck coefficient beta (ablation) | AENIB ResNet-18 · CIFAR-10-C | corruption error | Hurt |
| Remove the reconstruction loss (ablation) | AENIB ResNet-18 · CIFAR-10-C | corruption error | Hurt |
| Add the reconstruction loss (ablation) | AENIB ResNet-18 · CIFAR-10 | accuracy | Improved |
| Remove the nuisance loss (ablation) | AENIB ResNet-18 · CIFAR-10 | accuracy | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PixMix — PixMix: Dreamlike Pictures Comprehensively Improve Safety Measures (CVPR 2022) · better on 3 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- ROME is Forged in Adversity: Robust Distilled Datasets via Information Bottleneck (ICML 2025 · similarity 0.89)
- Distilling Robust and Non-Robust Features in Adversarial Examples by Information Bottleneck (NeurIPS 2021 · similarity 0.88)
- Maximum-Entropy Adversarial Data Augmentation for Improved Generalization and Robustness (NeurIPS 2020 · similarity 0.87)
- Adversarial Training Can Provably Improve Robustness: Theoretical Analysis of Feature Learning Process Under Structured Data (ICLR 2025 · similarity 0.86)
- DAT: Improving Adversarial Robustness via Generative Amplitude Mix-up in Frequency Domain (NeurIPS 2024 · similarity 0.86)
- Adversarial Robustness via Deformable Convolution with Stochasticity (ICML 2025 · similarity 0.86)
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.
