Offline Recommender System Evaluation under Unobserved Confounding

1 code implementation • 8 Sep 2023 • Olivier Jeunen, Ben London

Because the data collection policy is typically under the practitioner's control, the unconfoundedness assumption is often left implicit, and its violations are rarely dealt with in the existing literature.

Decision Making Recommendation Systems

Practical Bandits: An Industry Perspective

no code implementations • 2 Feb 2023 • Bram van den Akker, Olivier Jeunen, Ying Li, Ben London, Zahra Nazari, Devesh Parekh

The research literature on these topics is broad and vast, but this can overwhelm practitioners, whose primary aim is to solve practical problems, and therefore need to decide on a specific instantiation or approach for each project.

Decision Making Decision Making Under Uncertainty

Off-policy evaluation for learning-to-rank via interpolating the item-position model and the position-based model

no code implementations • 15 Oct 2022 • Alexander Buchholz, Ben London, Giuseppe Di Benedetto, Thorsten Joachims

A critical need for industrial recommender systems is the ability to evaluate recommendation policies offline, before deploying them to production.

Learning-To-Rank Off-policy evaluation +1

Boosted Off-Policy Learning

no code implementations • 1 Aug 2022 • Ben London, Levi Lu, Ted Sandler, Thorsten Joachims

We propose the first boosting algorithm for off-policy learning from logged bandit feedback.

Bayesian Counterfactual Risk Minimization

no code implementations • 29 Jun 2018 • Ben London, Ted Sandler

We present a Bayesian view of counterfactual risk minimization (CRM) for offline learning from logged bandit feedback.

A PAC-Bayesian Analysis of Randomized Learning with Application to Stochastic Gradient Descent

1 code implementation • NeurIPS 2017 • Ben London

This inspires an adaptive sampling algorithm for SGD that optimizes the posterior at runtime.

Generalization Bounds

Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction

no code implementations • 26 Sep 2013 • Stephen Bach, Bert Huang, Ben London, Lise Getoor

Graphical models for structured domains are powerful tools, but the computational complexities of combinatorial prediction spaces can force restrictions on models, or require approximate inference in order to be tractable.

Structured Prediction

Multi-relational Learning Using Weighted Tensor Decomposition with Modular Loss

no code implementations • 7 Mar 2013 • Ben London, Theodoros Rekatsinas, Bert Huang, Lise Getoor

For the typical cases of real-valued functions and binary relations, we propose several loss functions and derive the associated parameter gradients.

Relational Reasoning Tensor Decomposition

Graph-based Generalization Bounds for Learning Binary Relations

no code implementations • 21 Feb 2013 • Ben London, Bert Huang, Lise Getoor

We investigate the generalizability of learned binary relations: functions that map pairs of instances to a logical indicator.

Entity Resolution Generalization Bounds +1