Search Results for author:
Found 18 papers, 6 papers with code
Optimal Baseline Corrections for Off-Policy Contextual Bandits
The foundation of our framework is the derivation of an equivalent baseline correction for all of the existing control variates.
Multi-Objective Recommendation via Multivariate Policy Learning
We frame this as a decision-making task, where the scalarisation weights are actions taken to maximise an overall North Star reward (e. g. long-term user retention or growth).
Learning Metrics that Maximise Power for Accelerated A/B-Tests
Online controlled experiments are a crucial tool to allow for confident decision-making in technology companies.
Variance Reduction in Ratio Metrics for Efficient Online Experiments
The main culprit for this inefficiency is the variance of the online metrics.
Learning-to-Rank with Nested Feedback
Many platforms on the web present ranked lists of content to users, typically optimized for engagement-, satisfaction- or retention- driven metrics.
On Gradient Boosted Decision Trees and Neural Rankers: A Case-Study on Short-Video Recommendations at ShareChat
We highlight (1) neural networks' ability to handle large training data size, user- and item-embeddings allows for more accurate models than GBDTs in this setting, and (2) because GBDTs are less reliant on specialised hardware, they can provide an equally accurate model at a lower cost.
Ad-load Balancing via Off-policy Learning in a Content Marketplace
We start by presenting a motivating analysis of the ad-load balancing problem, highlighting the conflicting objectives between user satisfaction and ads revenue.
Offline Recommender System Evaluation under Unobserved Confounding
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.
On (Normalised) Discounted Cumulative Gain as an Off-Policy Evaluation Metric for Top-$n$ Recommendation
Through a correlation analysis between off- and on-line experiments conducted on a large-scale recommendation platform, we show that our unbiased DCG estimates strongly correlate with online reward, even when some of the metric's inherent assumptions are violated.
A Probabilistic Position Bias Model for Short-Video Recommendation Feeds
Empirical insights from a large-scale social media platform show how our probabilistic position bias model more accurately captures empirical exposure than existing models, and paves the way for unbiased evaluation and learning-to-rank.
RecFusion: A Binomial Diffusion Process for 1D Data for Recommendation
In this paper we propose RecFusion, which comprise a set of diffusion models for recommendation.
A Common Misassumption in Online Experiments with Machine Learning Models
Online experiments such as Randomised Controlled Trials (RCTs) or A/B-tests are the bread and butter of modern platforms on the web.
Practical Bandits: An Industry Perspective
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.
Disentangling Causal Effects from Sets of Interventions in the Presence of Unobserved Confounders
We address this challenge by aiming to learn the effect of a single-intervention from both observational data and sets of interventions.
Offline Evaluation of Reward-Optimizing Recommender Systems: The Case of Simulation
Naturally, the reason for this is that we can directly measure utility metrics that rely on interventions, being the recommendations that are being shown to users.
Learning from Bandit Feedback: An Overview of the State-of-the-art
In order to handle this "bandit-feedback" setting, several Counterfactual Risk Minimisation (CRM) methods have been proposed in recent years, that attempt to estimate the performance of different policies on historical data.
On the Value of Bandit Feedback for Offline Recommender System Evaluation
The procedure aims to give an answer to the question: "Given the natural sequence of user-item interactions up to time t, can we predict which item the user will interact with at time t+1?".
Three Methods for Training on Bandit Feedback
There are three quite distinct ways to train a machine learning model on recommender system logs.
