Leonardo Cella
YOU?
Author Swipe
View article: Divide-and-conquer with finite sample sizes: valid and efficient possibilistic inference
Divide-and-conquer with finite sample sizes: valid and efficient possibilistic inference Open
Divide-and-conquer methods use large-sample approximations to provide frequentist guarantees when each block of data is both small enough to facilitate efficient computation and large enough to support approximately valid inferences. When …
View article: Fusing independent inferential models in a black-box manner
Fusing independent inferential models in a black-box manner Open
Inferential models (IMs) represent a novel possibilistic approach for achieving provably valid statistical inference. This paper introduces a general framework for fusing independent IMs in a "black-box" manner, requiring no knowledge of t…
View article: Computationally efficient variational-like approximations of possibilistic inferential models
Computationally efficient variational-like approximations of possibilistic inferential models Open
Inferential models (IMs) offer provably reliable, data-driven, possibilistic statistical inference. But despite the IM framework's theoretical and foundational advantages, efficient computation is a challenge. This paper presents a simple …
View article: Possibility-theoretic statistical inference offers performance and probativeness assurances
Possibility-theoretic statistical inference offers performance and probativeness assurances Open
Statisticians are largely focused on developing methods that perform well in a frequentist sense -- even the Bayesians. But the widely-publicized replication crisis suggests that these performance guarantees alone are not enough to instill…
View article: Group Meritocratic Fairness in Linear Contextual Bandits
Group Meritocratic Fairness in Linear Contextual Bandits Open
We study the linear contextual bandit problem where an agent has to select one candidate from a pool and each candidate belongs to a sensitive group. In this setting, candidates' rewards may not be directly comparable between groups, for e…
View article: Meta Representation Learning with Contextual Linear Bandits
Meta Representation Learning with Contextual Linear Bandits Open
Meta-learning seeks to build algorithms that rapidly learn how to solve new learning problems based on previous experience. In this paper we investigate meta-learning in the setting of stochastic linear bandit tasks. We assume that the tas…
View article: Multi-task Representation Learning with Stochastic Linear Bandits
Multi-task Representation Learning with Stochastic Linear Bandits Open
We study the problem of transfer-learning in the setting of stochastic linear bandit tasks. We consider that a low dimensional linear representation is shared across the tasks, and study the benefit of learning this representation in the m…
View article: EFFICIENCY AND REALISM IN STOCHASTIC BANDITS
EFFICIENCY AND REALISM IN STOCHASTIC BANDITS Open
This manuscript is dedicated to the analysis of the application of stochastic bandits to the recommender systems domain. Here a learning agent sequentially recommends one item from a catalog of available alternatives. Consequently, the env…
View article: Online Model Selection: a Rested Bandit Formulation
Online Model Selection: a Rested Bandit Formulation Open
Motivated by a natural problem in online model selection with bandit information, we introduce and analyze a best arm identification problem in the rested bandit setting, wherein arm expected losses decrease with the number of times the ar…
View article: Meta-learning with Stochastic Linear Bandits
Meta-learning with Stochastic Linear Bandits Open
We investigate meta-learning procedures in the setting of stochastic linear bandits tasks. The goal is to select a learning algorithm which works well on average over a class of bandits tasks, that are sampled from a task-distribution. Ins…
View article: Strong validity, consonance, and conformal prediction
Strong validity, consonance, and conformal prediction Open
Valid prediction of future observations is an important and challenging problem. The two general approaches for quantifying uncertainty about the future value employ prediction regions and predictive distribution, respectively, with the la…
View article: Valid distribution-free inferential models for prediction
Valid distribution-free inferential models for prediction Open
A fundamental problem in statistics and machine learning is that of using observed data to predict future observations. This is particularly challenging for model-based approaches because often the goal is to carry out this prediction with…
View article: Stochastic Bandits with Delay-Dependent Payoffs
Stochastic Bandits with Delay-Dependent Payoffs Open
Motivated by recommendation problems in music streaming platforms, we propose a nonstationary stochastic bandit model in which the expected reward of an arm depends on the number of rounds that have passed since the arm was last pulled. Af…
View article: Efficient Linear Bandits through Matrix Sketching
Efficient Linear Bandits through Matrix Sketching Open
We prove that two popular linear contextual bandit algorithms, OFUL and Thompson Sampling, can be made efficient using Frequent Directions, a deterministic online sketching technique. More precisely, we show that a sketch of size $m$ allow…
View article: Efficient Context-Aware Sequential Recommender System
Efficient Context-Aware Sequential Recommender System Open
Traditional collaborative filtering, and content-based approaches attempt to learn a static recommendation model in a batch fashion. These approaches are not suitable in highly dynamic recommendation scenarios, like news recommendation and…
View article: Exploring the Semantic Gap for Movie Recommendations
Exploring the Semantic Gap for Movie Recommendations Open
In the last years, there has been much attention given to the semantic gap problem in multimedia retrieval systems. Much effort has been devoted to bridge this gap by building tools for the extraction of high-level, semantics-based feature…
View article: Deriving Item Features Relevance from Past User Interactions
Deriving Item Features Relevance from Past User Interactions Open
Item-based recommender systems suggest products based on the similarities between items computed either from past user preferences (collaborative filtering) or from item content features (content-based filtering). Collaborative filtering h…