Ben London

Latest

• Counterfactual Ranking Evaluation with Flexible Click Models
• Contextual Position Bias Estimation Using a Single Stochastic Logging Policy
• Double clipping: Less-biased variance reduction in off-policy evaluation
• Offline Recommender System Evaluation under Unobserved Confounding
• Self-Normalized Off-Policy Estimators for Ranking
• Boosted Off-Policy Learning
• Control Variate Diagnostics for Detecting Problems in Logged Bandit Feedback
• Off-Policy Evaluation for Learning-to-Rank via Interpolating the Item-Position Model and the Position-Based Model
• Recommendations as Treatments
• Offline Policy Evaluation with New Arms
• PAC-Identifiability in Federated Learning
• Bayesian Counterfactual Risk Minimization
• Train and Test Tightness of LP Relaxations in Structured Prediction
• Bayesian Counterfactual Risk Minimization
• Sustainability at Scale: Bridging the Intention-Behavior Gap with Sustainable Recommendations
• A PAC-Bayesian Analysis of Randomized Learning with Application to Stochastic Gradient Descent
• Bounding the Integrality Distance of LP Relaxations for Structured Prediction
• Generalization Bounds for Randomized Learning with Application to Stochastic Gradient Descent
• Generative Adversarial Structured Networks
• Stability and Generalization in Structured Prediction
• Budgeted Online Collective Inference
• Collective Graph Identification
• On the Stability of Structured Prediction
• Online Inference for Knowledge Graph Construction.
• The Benefits of Learning with Strongly Convex Approximate Inference
• On the Strong Convexity of Variational Inference
• PAC-Bayesian Collective Stability
• Collective Activity Detection using Hinge-loss Markov Random Fields
• Collective Classification of Network Data
• Collective Stability in Structured Prediction: Generalization from One Example
• Empirical Analysis of Collective Stability
• Graph-based Generalization Bounds for Learning Binary Relations
• Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction
• Multi-relational Learning Using Weighted Tensor Decomposition with Modular Loss
• PAC-Bayes Generalization Bounds for Randomized Structured Prediction
• Improved Generalization Bounds for Large-scale Structured Prediction
• Multi-relational Weighted Tensor Decomposition
• Query-driven Active Surveying for Collective Classification
• Reducing Label Cost by Combining Feature Labels and Crowdsourcing