Safety, Robustness & Interpretability
Adversarial robustness, AI safety, hallucination, interpretability/explainability.
A topic in the Science corpus, drawn from 2,034 papers.
Trusted voices in Safety, Robustness & Interpretability
Ordered by the trust Safety, Robustness & Interpretability's own network concentrates on each researcher, most first.
- Ian J. Goodfellow β trusted by 1,700 researchers here
- Rob Fergus β trusted by 1,274 researchers here
- Christian Szegedy β trusted by 1,098 researchers here
- Nicholas Carlini β trusted by 843 researchers here
- Matthew D. Zeiler β trusted by 1,070 researchers here
- Cynthia Dwork β trusted by 765 researchers here
- David A. Wagner β trusted by 760 researchers here
- Jonathon Shlens β trusted by 828 researchers here
- Nicolas Papernot β trusted by 649 researchers here
- Patrick Drew McDaniel β trusted by 602 researchers here
- Percy Liang β trusted by 666 researchers here
- Sameer Singh β trusted by 570 researchers here
Most-cited papers in Safety, Robustness & Interpretability
Ordered by the recorded citation count from the source corpus; a citation is not an endorsement.
- A Unified Approach to Interpreting Model Predictions (2017) β 37,902 citations
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization (2016) β 28,183 citations
- Explaining and Harnessing Adversarial Examples (2014) β 22,661 citations
- Visualizing and Understanding Convolutional Networks (2013) β 17,221 citations
- Intriguing properties of neural networks (2013) β 16,794 citations
- Towards Deep Learning Models Resistant to Adversarial Attacks (2017) β 15,165 citations
- Anomaly detection: A survey (2009) β 12,600 citations
- Towards Evaluating the Robustness of Neural Networks (2016) β 10,045 citations
- Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead (2018) β 9,528 citations
- k-Anonymity: A Model for Protecting Privacy (2002) β 9,049 citations
- Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, Information Fusion) β 9,028 citations
- On Calibration of Modern Neural Networks (2017) β 8,827 citations
- Deep Learning with Differential Privacy (2016) β 8,077 citations
- On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? π¦ (2021) β 7,969 citations
- Differential Privacy (2006) β 7,564 citations
- On the Opportunities and Risks of Foundation Models (2021) β 6,923 citations
- Adversarial examples in the physical world (2016) β 6,799 citations
- What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision? (2017) β 6,295 citations
- A Survey on Bias and Fairness in Machine Learning (2019, ACM Computing Surveys) β 6,091 citations
- Dissecting racial bias in an algorithm used to manage the health of populations (2019) β 5,573 citations
- DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks (2015) β 5,469 citations
- Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification (2018) β 5,467 citations
- Explanation in Artificial Intelligence: Insights from the Social Sciences (2017) β 5,450 citations
- Membership Inference Attacks Against Machine Learning Models (2016) β 5,446 citations
- Towards A Rigorous Science of Interpretable Machine Learning (2017) β 5,317 citations
- Equality of Opportunity in Supervised Learning (2016) β 5,258 citations
- Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI) (2018) β 5,195 citations
- Survey of Hallucination in Natural Language Generation (2022) β 5,175 citations
- A Survey of Methods for Explaining Black Box Models (2018) β 5,098 citations
- Striving for Simplicity: The All Convolutional Net (2014) β 5,049 citations
- On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation (2015) β 4,977 citations
- Learning Important Features Through Propagating Activation Differences (2017) β 4,716 citations
- Fairness through awareness (2011) β 4,555 citations
- The mythos of model interpretability (2016) β 4,511 citations
- A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks (2016) β 4,454 citations
- The Limitations of Deep Learning in Adversarial Settings (2015) β 4,317 citations
- Sorting Things Out: Classification and Its Consequences (2001) β 4,185 citations
- Differential Privacy: A Survey of Results (2008) β 4,153 citations
- Practical Black-Box Attacks against Machine Learning (2016) β 4,113 citations
- Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (2016) β 3,784 citations
- Understanding Black-box Predictions via Influence Functions (2017) β 3,706 citations
- Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples (2018) β 3,560 citations
- Deep neural networks are easily fooled: High confidence predictions for unrecognizable images (2014) β 3,543 citations
- Adversarial Machine Learning at Scale (2016) β 3,411 citations
- Boosting Adversarial Attacks with Momentum (2017) β 3,379 citations
- Constitutional AI: Harmlessness from AI Feedback (2022, arXiv.org) β 3,377 citations
- Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures (2015) β 3,376 citations
- Fake News Detection on Social Media: A Data Mining Perspective (2017) β 3,316 citations
- Distillation as a Defense to Adversarial Perturbations Against Deep Neural Networks (2015, IEEE Symposium on Security and Privacy) β 3,315 citations
- Concrete Problems in AI Safety (2016) β 3,295 citations
Most active researchers
By papers in this topic: a volume measure, not trust.
- Nicholas Carlini β 32 papers
- Nicolas Papernot β 30 papers
- Dawn Song β 26 papers
- Percy Liang β 26 papers
- Battista Biggio β 26 papers
- Ian J. Goodfellow β 24 papers
- Fabio Roli β 23 papers
- Kai-Wei Chang β 21 papers
- Bo Li β 20 papers
- Been Kim β 20 papers
- Cynthia Rudin β 20 papers
- Florian TramΓ¨r β 19 papers