Diversity regularization in deep ensembles
Explore this paper's citation graph
Summary
This work is proposing a strategy for training deep ensembles with a diversity function regularization, which improves the calibration property while maintaining a similar prediction accuracy.
- Type
- preprint
- Published
- 2018-02-22
- Cited by
- 16
- References
- 8
- Access
- Open access
- OpenAlex
- https://openalex.org/W2788453699
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:3650705
Keywords
Regularization (linguistics), Diversity (politics), Artificial intelligence, Statistical physics, Mathematics
References
- The Comparison and Evaluation of Forecasters.
- Ensemble learning via negative correlation
- Obtaining Well Calibrated Probabilities Using Bayesian Binning
- End to End Learning for Self-Driving Cars
- On Calibration of Modern Neural Networks
- Safe Visual Navigation via Deep Learning and Novelty Detection
- Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
- On Fairness and Calibration
Cited by
- Maximizing Overall Diversity for Improved Uncertainty Estimates in Deep Ensembles
- Machine learning can identify newly diagnosed patients with CLL at high risk of infection
- DICE: Diversity in Deep Ensembles via Conditional Redundancy Adversarial Estimation
- Ensembles for Uncertainty Estimation: Benefits of Prior Functions and Bootstrapping
- Ensemble of Pre-Trained Neural Networks for Segmentation and Quality Detection of Transmission Electron Microscopy Images
- Regularization Strength Impact on Neural Network Ensembles
- Bayesian Quadrature for Neural Ensemble Search
- Deep Anti-Regularized Ensembles provide reliable out-of-distribution uncertainty quantification
- Diversifying Deep Ensembles: A Saliency Map Approach for Enhanced OOD Detection, Calibration, and Accuracy
- MODL: Multilearner Online Deep Learning
- Decentralized Low-Latency Collaborative Inference via Ensembles on the Edge
- Improving Generalization of End-to-End ASR through Diversity and Independence Regularization
- Exploring the Rashomon Set for Concept-Based Models
- Verification-Aided Deep Ensemble Selection
- Diversifying Deep Ensembles: A Saliency Map Approach for Enhanced OOD Detection, Calibration, and Accuracy
- Improving robustness and calibration in ensembles with diversity regularization
Related papers
- Establishing Differences Between Diversity Requirements and Other Courses with Varying Degrees of Diversity Inclusivity
- CLOAKED CULTURE AND VEILED DIVERSITY: WHY THEORISTS IGNORED EARLY U.S. WORKFORCE DIVERSITY.
- Diversity discourses and the articulation of discrimination: the case of public organisations
- Increasing diversity in radiologic technology.
- 'Diversity Begets Diversity’: Polyarchy and Institutional Diversity for Plant Genetic Resources
- Praktik-Praktik Manajemen Keragaman di Tempat Kerja
- AC 2007-1966: DIVERSITY IN ENGINEERING EDUCATION RESEARCH: INSIGHTS FROM THREE STUDY DESIGNS
- Following diversity through the university: On knowing and embodying a problem
- Dealing with diversity in workgroups: Preventing problems and promoting potential