The field of machine learning and optimization is rapidly evolving, with a focus on developing innovative methods to address complex problems. Recent research has explored the application of bilevel optimization, Bayesian optimization, and online learning to various domains, including regression analysis, bilateral trade, and medical imaging. Noteworthy papers have proposed new approaches to countering adversarial evasion, learning in echo chambers, and evaluating temperature scaling calibration effectiveness. Additionally, researchers have made significant progress in developing new algorithms and frameworks for online decision-making, distributionally robust optimization, and private online learning. These advancements have the potential to improve the reliability and efficiency of machine learning models in real-world applications. Notable papers include: Information-Theoretic Bayesian Optimization for Bilevel Optimization Problems, which proposes an information-theoretic approach to Bayesian optimization for bilevel optimization problems. Countering adversarial evasion in regression analysis, which proposes a pessimistic bilevel optimization program for regression scenarios to mitigate threats to classifiers. Calibration Meets Reality: Making Machine Learning Predictions Trustworthy, which presents a rigorous theoretical analysis of post-hoc calibration methods and explores the impact of feature informativeness on calibration performance.
Advances in Machine Learning and Optimization
Sources
Machine Learning. The Science of Selection under Uncertainty
Information-Theoretic Bayesian Optimization for Bilevel Optimization Problems
Countering adversarial evasion in regression analysis
Nearly Tight Regret Bounds for Profit Maximization in Bilateral Trade
Calibration Meets Reality: Making Machine Learning Predictions Trustworthy
The Role of Commitment in Optimal Stopping
An SoS Entropy Dichotomy via Windowed Hypercontractivity
A Bilevel Approach to Integrated Surgeon Scheduling and Surgery Planning solved via Branch-and-Price
Evaluating Temperature Scaling Calibration Effectiveness for CNNs under Varying Noise Levels in Brain Tumour Detection
Learning in an Echo Chamber: Online Learning with Replay Adversary
Online Decision Making with Generative Action Sets
Iterative Hypothesis Pruning and Distribution-based Early Labeling for Sequential Hypothesis Testing
From Fragile to Certified: Wasserstein Audits of Group Fairness Under Distribution Shift
Wasserstein Distributionally Robust Optimization Through the Lens of Structural Causal Models and Individual Fairness
Approximately Unimodal Likelihood Models for Ordinal Regression
Lipschitz Bandits with Stochastic Delayed Feedback
Private Online Learning against an Adaptive Adversary: Realizable and Agnostic Settings
Designing Ambiguity Sets for Distributionally Robust Optimization Using Structural Causal Optimal Transport
The Good, the Bad, and the Sampled: a No-Regret Approach to Safe Online Classification
PASTA: A Unified Framework for Offline Assortment Learning
Universal Dynamic Regret and Constraint Violation Bounds for Constrained Online Convex Optimization
Lower Bounds on Adversarial Robustness for Multiclass Classification with General Loss Functions