arXiv subjects

Martin Bompaire

Publications and source records attributed to Martin Bompaire.

7 recordsLinked to original sources

arXiv2022

Offline Evaluation of Reward-Optimizing Recommender Systems: The Case of Simulation

Both in academic and industry-based research, online evaluation methods are seen as the golden standard for interactive applications like recommendation systems. 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. Nevertheless, online evaluation methods are costly for a number of reasons, and a clear need remains for reliable offline evaluation procedures. In industry, offline metrics are often used as a first-line evaluation to generate promising candidate models to evaluate online. In academic work, limited access to online systems makes offline metrics the de facto approach to validating novel methods. Two classes of offline metrics exist: proxy-based methods, and counterfactual methods. The first class is often poorly correlated with the online metrics we care about, and the latter class only provides theoretical guarantees under assumptions that cannot be fulfilled in real-world environments. Here, we make the case that simulation-based comparisons provide ways forward beyond offline metrics, and argue that they are a preferable means of evaluation.

cs.IR↗

arXiv2020

Fixed point label attribution for real-time bidding

Problem definition: Most of the display advertising inventory is sold through real-time auctions. The participants of these auctions are typically bidders (Google, Criteo, RTB House, Trade Desk for instance) who participate on behalf of advertisers. In order to estimate the value of each display opportunity, they usually train advanced machine learning algorithms using historical data. In the labeled training set, the inputs are vectors of features representing each display opportunity and the labels are the generated rewards. In practice, the rewards are given by the advertiser and are tied to whether or not a particular user converts. Consequently, the rewards are aggregated at the user level and never observed at the display level. A fundamental task that has, to the best of our knowledge, been overlooked is to account for this mismatch and split, or attribute, the rewards at the right granularity level before training a learning algorithm. We call this the label attribution problem. Methodology/results: In this paper, we develop an approach to the label attribution problem, which is both theoretically justified and practical. In particular, we develop a fixed point algorithm that allows for large scale implementation and showcase our solution using a large scale publicly available dataset from Criteo, a large Demand Side Platform. We dub our approach the Fixed Point Label Attribution (FiPLA) Algorithm. Managerial implications: There is often a hidden leap of faith when transforming the advertiser's signal into display labelling. DSP providers should be careful when building their machine learning pipeline and carefully solve the label attribution step.

cs.GT↗

arXiv2020

Sparse and low-rank multivariate Hawkes processes

We consider the problem of unveiling the implicit network structure of node interactions (such as user interactions in a social network), based only on high-frequency timestamps. Our inference is based on the minimization of the least-squares loss associated with a multivariate Hawkes model, penalized by $\ell_1$ and trace norm of the interaction tensor. We provide a first theoretical analysis for this problem, that includes sparsity and low-rank inducing penalizations. This result involves a new data-driven concentration inequality for matrix martingales in continuous time with observable variance, which is a result of independent interest and a broad range of possible applications since it extends to matrix martingales former results restricted to the scalar case. A consequence of our analysis is the construction of sharply tuned $\ell_1$ and trace-norm penalizations, that leads to a data-driven scaling of the variability of information available for each users. Numerical experiments illustrate the significant improvements achieved by the use of such data-driven penalizations.

stat.ML↗

arXiv2019

Learning from Bandit Feedback: An Overview of the State-of-the-art

In machine learning we often try to optimise a decision rule that would have worked well over a historical dataset; this is the so called empirical risk minimisation principle. In the context of learning from recommender system logs, applying this principle becomes a problem because we do not have available the reward of decisions we did not do. 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. Through importance sampling and various variance reduction techniques, these methods allow more robust learning and inference than classical approaches. It is difficult to accurately estimate the performance of policies that frequently perform actions that were infrequently done in the past and a number of different types of estimators have been proposed. In this paper, we review several methods, based on different off-policy estimators, for learning from bandit feedback. We discuss key differences and commonalities among existing approaches, and compare their empirical performance on the RecoGym simulation environment. To the best of our knowledge, this work is the first comparison study for bandit algorithms in a recommender system setting.

cs.IR↗

arXiv2019

Three Methods for Training on Bandit Feedback

There are three quite distinct ways to train a machine learning model on recommender system logs. The first method is to model the reward prediction for each possible recommendation to the user, at the scoring time the best recommendation is found by computing an argmax over the personalized recommendations. This method obeys principles such as the conditionality principle and the likelihood principle. A second method is useful when the model does not fit reality and underfits. In this case, we can use the fact that we know the distribution of historical recommendations (concentrated on previously identified good actions with some exploration) to adjust the errors in the fit to be evenly distributed over all actions. Finally, the inverse propensity score can be used to produce an estimate of the decision rules expected performance. The latter two methods violate the conditionality and likelihood principle but are shown to have good performance in certain settings. In this paper we review the literature around this fundamental, yet often overlooked choice and do some experiments using the RecoGym simulation environment.

cs.IR↗

arXiv2018

Dual optimization for convex constrained objectives without the gradient-Lipschitz assumption

The minimization of convex objectives coming from linear supervised learning problems, such as penalized generalized linear models, can be formulated as finite sums of convex functions. For such problems, a large set of stochastic first-order solvers based on the idea of variance reduction are available and combine both computational efficiency and sound theoretical guarantees (linear convergence rates). Such rates are obtained under both gradient-Lipschitz and strong convexity assumptions. Motivated by learning problems that do not meet the gradient-Lipschitz assumption, such as linear Poisson regression, we work under another smoothness assumption, and obtain a linear convergence rate for a shifted version of Stochastic Dual Coordinate Ascent (SDCA) that improves the current state-of-the-art. Our motivation for considering a solver working on the Fenchel-dual problem comes from the fact that such objectives include many linear constraints, that are easier to deal with in the dual. Our approach and theoretical findings are validated on several datasets, for Poisson regression and another objective coming from the negative log-likelihood of the Hawkes process, which is a family of models which proves extremely useful for the modeling of information propagation in social networks and causality inference.

stat.ML↗

arXiv2018

Tick: a Python library for statistical learning, with a particular emphasis on time-dependent modelling

Tick is a statistical learning library for Python~3, with a particular emphasis on time-dependent models, such as point processes, and tools for generalized linear models and survival analysis. The core of the library is an optimization module providing model computational classes, solvers and proximal operators for regularization. tick relies on a C++ implementation and state-of-the-art optimization algorithms to provide very fast computations in a single node multi-core setting. Source code and documentation can be downloaded from https://github.com/X-DataInitiative/tick

stat.ML↗