Enseignement

Filtrer

Enseignement

Algo-Rhythm Unplugged: Effects of explaining algorithmic recommendations on music discovery

Date : Jeudi | 2024-10-17 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Mehdi LOUAFI (LEO, UniversitĂ© d’OrlĂ©ans)

Music streaming services make massive use of algorithms in their music recommender systems (MRS) to guide users to tracks that they are likely to enjoy. However, the black-box nature of these algorithms makes them difficult for users to understand, both in terms of how they work and the music they predict. The field of explainable AI (XAI), and in particular its “explanation” side, has emerged to make the uses of AI (including MRSs) more comprehensible to users. This paper aims to observe, using a laboratory experiment, whether the explanation of an MRS algorithm induces a change in discovery behavior on music streaming services. In a theoretical framework, we model two types of user discovery behavior, namely “study” and “browse” behaviors. We then test in the lab the explanation effects of these behaviors by explaining a simplified “semi-personalized” MRS and measuring the relative listening time of the tracks. Data shows no average effect induced by the explanations, but we observe a differentiated impact of explanations based on the treatment intensity (i.e. the time spent reading them). The more individuals are treated, the more they listen to the tracks, reinforcing the “study” behavior.

First-year PhD. Students Presentation (2/2)

Date : Jeudi | 2024-10-10 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire du LÉO

 

S P E C I A L       S E S S I O N

 

Three new PhD. Students from the LEO (Université d'Orléans) will present their thesis projects. Each of them will give a 15 minutes presentation followed by a 15 minutes discussion :

  • Martin CimetiĂšre (Macro/Finance Team)
  • Marie Jezequel-Royer (International Economics/Sustainable Development Team)
  • Elhadji Mamadou Guene (Econometrics Team)

 

First-year PhD. Students Presentation (1/2)

Date : Jeudi | 2024-10-03 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire du LÉO

 

S P E C I A L       S E S S I O N

 

Two new PhD. Students from the LEO (Université d'Orléans) will present their thesis projects. Each of them will give a 15 minutes presentation followed by a 15 minutes discussion :

  • AĂŻssata Ousmane Dicko (International Economics/Sustainable Development Team)
  • Saniya Ali Zahed (International Economics/Sustainable Development Team)

 

Do Concerns for Risks of Personal Data Leaks Impact Decision-Making in Finance?

Date : Jeudi | 2024-05-23 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Mehdi LOUAFI (LEO, UniversitĂ© d’OrlĂ©ans)

Based on an experiment with participants from the general population, we examine the determinants of individuals' investment decisions in the presence of potential personal data leaks. We find that participants when presented with scenarios explicitly acknowledging the probability distribution and content of personal data leaks, tend to prioritize financial performance over the safeguarding of their personal data. However, we show that individuals adjust their preferences in favor of privacy when confronted with high risks of data leaks or low financial returns. Additionally, we observe that the magnitude of data leaks in previous rounds, as measured by the number of participants affected, decreases the likelihood of investing in options carrying privacy risks. While past exposures to data leaks at the aggregate level influence decision-making, individual experiences of data leaks in previous rounds do not significantly impact subsequent investment decisions. These findings contribute to our understanding of how individuals navigate complex trade-offs associated with privacy, shedding light on decision-making within the context of data security concerns and financial decisions.

Digital wage payments and financial inclusion in Developing and Emerging Economies

Date : Vendredi | 2024-04-26 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Ismail BELLO (LEO, UniversitĂ© d’OrlĂ©ans)

The study examines the impact of wage payment mode on financial inclusion in Developing and Emerging Economies. In such countries, non-bank financial institutions like mobile money agents have an important role as payment service providers. To account for this specificity, we build two indices: a traditional (banking) financial inclusion index, and a digital financial inclusion index. We test the hypothesis that distributing wages on financial accounts, through mobile phones or cards (which we label "digital wage payment"), does increase financial inclusion, whereas wages paid in cash do not. Countries wishing to improve their financial inclusion would therefore have a lever for action by regulating how salaries are paid.

Beyond the Classroom? Secondary School Expansion and Women’s Marital Outcomes in Senegal

Date : Jeudi | 2024-04-25 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Daniel PEREZ PARRA (ERUDITE, Université Gustave Eiffel)

The importance of women's education for the pursuit of gender equality in many facets of society is well-documented. I investigate the implications of a large-scale increase in secondary school capacity for women's education outcomes and marriage. Using the ambitious expansion of new classrooms in Senegal to test its impact on education and women's marital experiences, I find substantial gains in female education and learning indicators. However, I do not detect any significant shifts in child marriage, polygyny, or intimate partner violence overall. This analysis indicates the complexity of schooling policies in combating broader trends and suggests a broader strategy to affect multiple facets of social change.

Analyse des impacts du bruit routier sur la consommation mĂ©dicamenteuse dans l’étude Brouhaha de la cohorte de santĂ© E3N

Date : Jeudi | 2024-04-18 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Julie LEDUC (LEO, Université de Tours)

L'impact néfaste du bruit généré par les transports sur la santé est bien documenté, avec une attention particuliÚre portée aux effets cardio-métaboliques et psychologiques. Malgré une abondante littérature scientifique, des zones d'ombre persistent concernant certains aspects de cette problématique sur la santé. Dans cette étude, nous nous concentrons sur l'association entre le bruit routier et les différentes classes de médicaments selon la classification ATC (Anatomical Therapeutic Chemical), qui classe les principes actifs en fonction de leur cible thérapeutique selon cinq niveaux de granularité.
Notre Ă©tude transversale porte sur une population de 22 000 femmes rĂ©sidant en Île-de-France (IdF) ou en Auvergne-RhĂŽne-Alpes (AuRa) en 2014. Elle s'appuie sur les donnĂ©es de la cohorte prospective E3N (Étude ÉpidĂ©miologique auprĂšs des femmes de l'Éducation Nationale), composĂ©e initialement de 98 995 femmes françaises ĂągĂ©es de 40 Ă  65 ans et affiliĂ©es Ă  la Mutuelle GĂ©nĂ©rale de l'Éducation Nationale (MGEN) en 1990. Le gĂ©ocodage des adresses des participantes a Ă©tĂ© rĂ©alisĂ© Ă  l'aide du gĂ©ocodeur BD Adresse pour ArcGIs. L'exposition au bruit des transports des participantes a Ă©tĂ© Ă©valuĂ©e Ă  partir des Cartes StratĂ©giques de Bruit (CSB) et calculĂ©e en Lden (Niveau jour-soirĂ©e-nuit) pour les trois sources de bruit de transport : routier, ferroviaire et aĂ©rien. Les modĂšles ont Ă©tĂ© ajustĂ©s en tenant compte de la pollution atmosphĂ©rique par les particules fines de diamĂštre < 2,5 ÎŒm [PM2,5] et le dioxyde d'azote [NO2]. Nous avons Ă©tudiĂ© six modĂšles avec diffĂ©rents niveaux d'ajustement, incluant les caractĂ©ristiques personnelles (Ăąge + IMC), les habitudes de consommation (tabagisme, alcool) et les habitudes de vie (apport calorique, activitĂ© physique). Chaque individu a Ă©tĂ© classĂ© dans un groupe d'exposition en dĂ©cibels correspondant aux seuils de risques sanitaires dĂ©finis par l'Organisation Mondiale de la SantĂ©. Nous avons utilisĂ© un modĂšle logit en crĂ©ant une variable factice prenant la valeur 1 si la femme avait consommĂ© la molĂ©cule de chaque classe ATC. Les rapports de cotes estimĂ©s (OR), reflĂ©tant le risque de consommer la classe de mĂ©dicament en fonction de l'exposition au bruit, ont Ă©tĂ© calculĂ©s avec leurs intervalles de confiance associĂ©s (IC Ă  95%) pour chaque catĂ©gorie.
Nous avons observé une association positive entre l'exposition au bruit routier et le risque de consommer certaines classes de médicaments. Les détails de ces résultats seront présentés lors du séminaire ; ils viennent corroborer les conclusions de la littérature existante et ouvrent de nouvelles voies de recherche sur des effets de santé encore peu explorés. Cette étude souligne l'importance d'une compréhension approfondie des effets sur la santé de cette pollution, étant donné que plus de 50% des Européens sont exposés au bruit des transports. Le deuxiÚme volet de cette étude met également en lumiÚre les coûts sanitaires qui y sont associés, soulignant ainsi le besoin de politiques de santé publique efficaces pour réinternaliser ces coûts.

Poor Planning: Investors who Save Regularly Save Less Overall

Date : Jeudi | 2024-03-28 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Indigo JONES (LEO, UniversitĂ© d’OrlĂ©ans)

Scheduling regular deposits is consensus popular financial advice, but does doing so actually increase savings? Perhaps not. In a dataset of 36,836 robo-advisor clients who saved an average of €21,602, investors who scheduled deposits saved 32.6% (€8,630) less overall than investors who did not schedule deposits, and still 16.0% (€3,466) less than average ceteris paribus. Yet, investors who started or stopped scheduling deposits saved €3,247 more during those years in which they were scheduling deposits. Additionally, many factors such as an investor's age, sex, number of children, income, wealth, investment horizon, ESG status, investment goal, and financial literacy were significantly related to an investor's decision whether to make scheduled deposits and/or how much they eventually saved.

Factors Influencing the Decline of Manufacturing Pollution in the European Union: A Study of Productivity, Environmental Regulations, Expenditure, and Trade Costs

Date : Jeudi | 2024-03-11 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Sahar AMIDI (LEO, UniversitĂ© d’OrlĂ©ans)

The paper explores how productivity, expenditure share, trade cost, and environmental regulations affect pollution levels in Europe's manufacturing industry. The World Input-Output Database provides data on global and local pollution for each industrial sector solely for the period ranging from 1995 to 2009. We use a general equilibrium model and quantitative trade model that considers pollution as a byproduct of production. The study aims to examine the effectiveness of regulations and control for the primary causes of environmental pollution (the main causes). Our empirical results reveal that air pollution emissions from EU manufacturing decreased by 33.21 percent despite an 85.44 percent increase in real manufacturing output. This outcome could provide evidence for the role of reducing the pollution contamination of manufacturing. The study finds that most of the decrease in emissions can be ascribed to changes in environmental regulations, rather than changes in expenditure share, trade cost, and productivity. Increasing environmental regulations by 20 percent can eliminate emissions intensity. After increasing environmental regulations by 20%, the emission of global pollutants such as methane decreased by 17.27% in 2009.

Working Conditions and Mental Health

Date : Jeudi | 2023-02-15 Ă  12h30
Lieu : Salle des thĂšses

Lien TEAMS : Cliquer ici pour rejoindre le sĂ©minaire doctorant du LÉO

Aynur KHALILZADEH (LEO, UniversitĂ© d’OrlĂ©ans)

Work significantly influences individual well-being and autonomy, with growing acknowledgment of the detrimental effects of poor working conditions on physical and mental health. Psychosocial factors, encompassing both protective resources and risk factors, play a pivotal role in shaping mental and emotional well-being within the social environment. Despite global recognition of these risks, understanding their relative importance and interaction with individual characteristics remains limited. Leveraging data from the French Monitoring of Employees' Exposure to Occupational Risks 2017 survey (SUMER 2017), our research investigates the complex interplay between working conditions and mental health among 20,605 employees in France. Utilizing scientifically validated questionnaires, including the PHQ-9, we assess workers' mental health and exposure to psychosocial risks. Our findings reveal a significant association between exposure to psychosocial risks and depressive disorders, with logistic regression analysis highlighting the predictive power of these risks on mental health outcomes. Moreover, our study identifies gender and occupational disparities in psychosocial risk exposure, emphasizing the need for tailored interventions to promote worker well-being. This comprehensive analysis enhances our understanding of the nuanced relationships between psychosocial factors, mental health outcomes, and their distribution across diverse demographic and occupational dimensions.