Gaël Poux-Médard

Gaël Poux-Médard
Université de Lyon · Department of Computer Science

About

28
Publications
1,100
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
73
Citations
Introduction
My interests in research revolve around computational social sciences. I am currently exploring the role of information interactions in human choice. I apply my findings in this new branch of research to various real-world datasets --prediction of disease diagnosis, retweet on Twitter, human choice in dilemma contexts, Spotify playlist building, ... The methods I use range from stochastic block models to survival theory, which up to now have both proven their usefulness in modeling interactions.
Additional affiliations
October 2019 - present
Université de Lyon
Position
  • PhD Student
Education
September 2017 - September 2019
École Normale Supérieure de Lyon
Field of study
  • Physics, Complex systems, Digital humanities
September 2014 - September 2017

Publications

Publications (28)
Chapter
Full-text available
The Dirichlet process is one of the most widely used priors in Bayesian clustering. This process allows for a nonparametric estimation of the number of clusters when partitioning datasets. The “rich-get-richer” property is a key feature of this process, and transcribes that the a priori probability for a cluster to get selected dependent linearly o...
Chapter
The publication time of a document carries a relevant information about its semantic content. The Dirichlet-Hawkes process has been proposed to jointly model textual information and publication dynamics. This approach has been used with success in several recent works, and extended to tackle specific challenging problems –typically for short texts...
Chapter
Full-text available
Information spread on networks can be efficiently modeled by considering three features: documents’ content, time of publication relative to other publications, and position of the spreader in the network. Most previous works model up to two of those jointly, or rely on heavily parametric approaches. Building on recent Dirichlet-Point processes lit...
Chapter
Most models of information diffusion online rely on the assumption that pieces of information spread independently from each other. However, several works pointed out the necessity of investigating the role of interactions in real-world processes, and highlighted possible difficulties in doing so: interactions are sparse and brief. As an answer, re...
Preprint
Full-text available
The publication time of a document carries a relevant information about its semantic content. The Dirichlet-Hawkes process has been proposed to jointly model textual information and publication dynamics. This approach has been used with success in several recent works, and extended to tackle specific challenging problems --typically for short texts...
Preprint
Full-text available
Information spread on networks can be efficiently modeled by considering three features: documents' content, time of publication relative to other publications, and position of the spreader in the network. Most previous works model up to two of those jointly, or rely on heavily parametric approaches. Building on recent Dirichlet-Point processes lit...
Preprint
Full-text available
Since the development of writing 5000 years ago, human-generated data gets produced at an ever-increasing pace. Classical archival methods aimed at easing information retrieval. Nowadays, archiving is not enough anymore. The amount of data that gets generated daily is beyond human comprehension, and appeals for new information retrieval strategies....
Preprint
Full-text available
Last years have seen a regain of interest for the use of stochastic block modeling (SBM) in recommender systems. These models are seen as a flexible alternative to tensor decomposition techniques that are able to handle labeled data. Recent works proposed to tackle discrete recommendation problems via SBMs by considering larger contexts as input da...
Preprint
Full-text available
Most models of information diffusion online rely on the assumption that pieces of information spread independently from each other. However, several works pointed out the necessity of investigating the role of interactions in real-world processes, and highlighted possible difficulties in doing so: interactions are sparse and brief. As an answer, re...
Thesis
Full-text available
Since the development of writing 5000 years ago, human-generated data gets produced at an ever-increasing pace. This rate has been greatly influenced by technical innovations, such as clay tablets, papyrus, paper, press, and more recently the Internet. At the same time, new methods designed to handle and archive these growing information flows emer...
Article
Full-text available
The textual content of a document and its publication date are intertwined. For example, the publication of a news article on a topic is influenced by previous publications on similar issues, according to underlying temporal dynamics. However, it can be challenging to retrieve meaningful information when textual information conveys little informati...
Preprint
Full-text available
Large quantities of data flow on the internet. When a user decides to help the spread of a piece of information (by retweeting, liking, posting content), most research works assumes she does so according to information's content, publication date, the user's position in the network, the platform used, etc. However, there is another aspect that has...
Preprint
Full-text available
The textual content of a document and its publication date are intertwined. For example, the publication of a news article on a topic is influenced by previous publications on similar issues, according to underlying temporal dynamics. However, it can be challenging to retrieve meaningful information when textual information conveys little. Furtherm...
Article
Full-text available
Many studies have shown that there are regularities in the way human beings make decisions. However, our ability to obtain models that capture such regularities and can accurately predict unobserved decisions is still limited. We tackle this problem in the context of individuals who are given information relative to the evolution of market prices a...
Preprint
Full-text available
The textual content of a document and its publication date are intertwined. For example, the publication of a news article on a topic is influenced by previous publications on similar issues, according to underlying temporal dynamics. However, it can be challenging to retrieve meaningful information when textual information conveys little informati...
Conference Paper
Full-text available
In most real-world applications, it is seldom the case that a result appears independently from an environment. In social networks, users’ behavior results from the people they interact with, news in their feed, or trending topics. In natural language, the meaning of phrases emerges from the combination of words. In general medicine, a diagnosis is...
Conference Paper
Full-text available
Interactions between pieces of information (entities) play a substantial role in the way an individual acts on them: adoption of a product, the spread of news, strategy choice, etc. However, the underlying interaction mechanisms are often unknown and have been little explored in the literature. We introduce an efficient method to infer both the ent...
Preprint
Full-text available
Interactions between pieces of information (entities) play a substantial role in the way an individual acts on them: adoption of a product, the spread of news, strategy choice, etc. However, the underlying interaction mechanisms are often unknown and have been little explored in the literature. We introduce an efficient method to infer both the ent...
Preprint
Full-text available
Interactions between pieces of information (entities) play a substantial role in the way an individual acts on them: adoption of a product, the spread of news, strategy choice, etc. However, the underlying interaction mechanisms are often unknown and have been little explored in the literature. We introduce an efficient method to infer both the ent...
Preprint
Full-text available
One of the most used priors in Bayesian clustering is the Dirichlet prior. It can be expressed as a Chinese Restaurant Process. This process allows nonparametric estimation of the number of clusters when partitioning datasets. Its key feature is the "rich-get-richer" property, which assumes a cluster has an a priori probability to get chosen linear...
Preprint
Full-text available
Many studies have shown that there are regularities in the way human beings make decisions. However, our ability to obtain models that capture such regularities and can accurately predict unobserved decisions is still limited. We tackle this problem in the context of individuals who are given information relative to the evolution of market prices a...
Article
Full-text available
The identification of which nodes are optimal seeds for spreading processes on a network is a nontrivial problem that has attracted much interest recently. While activity has mostly focused on the nonrecurrent type of dynamics, here we consider the problem for the susceptible-infected-susceptible (SIS) spreading model, where an outbreak seeded in o...
Preprint
Full-text available
In most real-world applications, it is seldom the case that a given observable evolves independently of its environment. In social networks, users' behavior results from the people they interact with, news in their feed, or trending topics. In natural language, the meaning of phrases emerges from the combination of words. In general medicine, a dia...
Preprint
Full-text available
The identification of which nodes are optimal seeds for spreading processes on a network is a non-trivial problem that has attracted much interest recently. While activity has mostly focused on non-recurrent type of dynamics, here we consider the problem for the Susceptible-Infected-Susceptible (SIS) spreading model, where an outbreak seeded in one...

Network

Cited By