Coauteurs : Emanuele Campiglio (Universita di Bologna) et Davide Romelli (Trinity College Dublin)

Résumé : This paper studies the evolution and drivers of central bank communication on climate-related matters. We build a novel dataset containing 31,249 speeches from 131 central banks, over the 1986-2021 period. Using natural language processing techniques, we identify climate-related discourses and analyse their thematic content. We show how the rapid, but differentiated increase in climate-related communication can be linked to the rise of two distinct narratives, one centred around `green finance' as an opportunity, and the other around the threat represented by `climate-related risks'. We then study the drivers of those strands of climate communication and find that institutional dimensions - most prominently, the degree of involvement of central banks in financial supervision - are more significant in explaining climate-related communication than domestic exposure to physical and transition risks.

Résumé : Connue sous le nom de "ville des Sassi", Matera a fait l'objet d'un processus de rénovation impliquant toutes les régions de la Basilicate au cours des dernières années. Les ressources touristiques de la ville étaient pratiquement inconnues au niveau national et international, bien que les Sassi soient inscrits sur la liste du patrimoine mondial de l'UNESCO depuis 1993. La nomination de la Capitale européenne de la culture 2019 a déclenché une régénération intense, ouvrant la ville au tourisme mondial et révélant une grande résilience. Les expériences et les opinions des touristes ont été des ressources précieuses pour concevoir des activités touristiques et créer une nouvelle identité symbolique pour la ville, en particulier à l'ère du Web 2.0. Nous proposons ici de calculer les scores de polarité des avis et de les utiliser avec d'autres caractéristiques (par exemple, le prix, les services offerts et le type d'installations touristiques) pour créer des grappes spatiales selon la logique des indicateurs d'association spatiale locale (LISA). L'orientation sémantique géoréférencée des avis concernant une activité ou une attraction particulière représente une caractéristique quantitative utile pour des analyses ultérieures et la production de statistiques territoriales. La proposition peut être étendue à d'autres cas pour suivre l'évolution des sentiments à l'égard de domaines d'intérêt spécifiques et planifier d'éventuelles politiques d'intervention.

Résumé : Travaux appliqués du département recherche d’OpenStudio en IA et Économie : Conception d'outils de modélisation et de simulation des transformations des chaînes de valeur industrielles et d’accompagnement des entreprises pour aller vers des filières industrielles européennes plus autonomes, agiles et résilientes.

https://atlas.productive-synergies.com.

Résumé : Recently, Artificial Intelligence (AI) technology use has been rising in sports. For example, to reduce staff during the COVID-19 pandemic, major tennis tournaments replaced human line judges with Hawk-Eye Live technology. AI is now ready to move beyond such mundane tasks, however. A case in point and a perfect application ground is chess. To reduce the growing incidence of draws, many elite tournaments have resorted to fast chess tiebreakers. However, these tiebreakers are vulnerable to strategic manipulation, e.g., in the last game of the 2018 World Chess Championship, Carlsen—in a significantly advantageous position—offered a draw to Caruana (whom accepted the offer) to proceed to fast chess tiebreaks in which Carlsen had even better odds of winning the championship. By contrast, we prove that our AI-based method can serve as a judge to break ties without being vulnerable to such manipulation. It relies on measuring the difference between the evaluations of a player's actual move and the best move as deemed by a powerful chess engine. If there is a tie, the player with the higher quality measure wins the tiebreak. We generalize our method to all competitive sports and games in which AI's superiority is—or can be—established.

Abstract:

Around the world, countless workers perform data-driven tasks on online labour platforms to fuel the digital economy. Mostly brief, repetitive, and poorly paid, these so-called ‘micro-tasks’ include, for example, tagging objects in images, recording videos, and transcribing text, mostly for the artificial intelligence industry. Although these platforms have been harshly criticized for precarious working conditions and low wages, access to them is easy and open, even to unskilled people. In principle, women with care duties can be expected to benefit from flexible working hours and the possibility of working from home. However, this new form of online work fails to bridge the digital gender gap and may even exacerbate it. I demonstrate this result in three steps. First, inherited inequalities in the professional and domestic spheres turn platform-mediated micro-tasks into a ‘third shift’ on top of already burdened schedules. Second, the human capital of male and female data workers differs insofar as women are less likely to have received training in scientific and technological fields. Third, their social capital differs: using a ‘position generator’, a specific tool to capture workers' access to information and support resources that can come from their contacts with people in different occupations, I show that women have fewer ties to digital-related professionals who could provide them with knowledge and advice on how to successfully navigate the world of platforms. Taken together, these factors leave women with fewer career prospects in a technology-driven workforce and reproduce their relegation to lower-level computer jobs already observed in the early history of 20th century technology.

Rapporteur : Bastien Lextrait

Résumé : The ability to identify early-stage venture potential, which has traditionally relied on experts, is challenging because new ideas are uncertain. However, judgment from a large number of evaluators can effectively identify promising ventures. This paper empirically studies whether crowd-based information generates informative feedback for entrepreneurs and seed investors. Using data on 701 early-stage founders of new ventures examined by 2,600 evaluators, I estimated the effects of crowd-based ratings on survival and securing seed funds. I find that the crowd rating was uncorrelated with subsequent financing events from seed investors but increased a venture’s probability of continuation by 3.1 percentage points. This effect became stronger as the number of evaluators increased. This finding highlights that a judgment aggregation offers information for nascent venture founders.

Rapporteur : Mathieu Bernard
Résumé : Il peut arriver qu'un territoire soit sous l'emprise d'un pouvoir illégitime, contre lequel une partie de la population souhaite résister. De même, une population peut vouloir attirer l'attention d'un gouvernement sourd à ses revendications. Un des principaux moyens est alors la perturbation des infrastructures installées sur le territoire, typiquement les transports (routiers, aériens, maritimes), l'approvisionnement en ressources (énergie, eau), ou les communications (internet, médias). La plupart de ces infrastructures reposent sur des réseaux ; il s'agit donc d'y trouver des points faibles, de les bloquer ou de les saturer. Pour ce faire, plusieurs stratégies sont envisageables. Elles posent des questions de recherche en informatique que je propose de discuter dans cet exposé.

Rapporteure : Nadine Levratto

Rapporteur : César Ducruet

Article à l'appui de la présentation : Le solaire décentralisé à l’assaut des villes africaines Une analyse originale d’imagerie satellite et de Deep Learning

Rapporteur : Christophe Boucher

Rapporteur : Eric Darmon

Article à l'appui de la présentation : Artificial intelligence and consumer manipulations: from consumer's counter algorithms to firm's self‑regulation tools