At the Origins of Learning: Absorbing Knowledge Flows from within the Team

Charles Chadi Ayoubi, Michele Pezzoni, Fabiana Visentin
Elsevier, 2017
Article

Résumé

Empirical studies document a positive effect of collaboration on team productivity. However, little has been done to assess how knowledge flows among team members. Our study addresses this issue by exploring unique rich data on a Swiss funding program promoting research team collaboration. We find that being involved in an established collaboration and team size foster the probability of an individual learning from the other team members. We also find that team members with limited experience are more likely to learn from experienced peers. Moreover, there is an inverted U-shaped effect of cognitive distance on the probability of learning from other team members.

Official source

https://infoscience.epfl.ch/record/227509?ln=fr

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

Publications associées (13)

This thesis presents four essays providing novel empirical and theoretical insights on the incentives and institutional structures that favor knowledge production and diffusion. The first two studies analyze these processes in the realm of scientific research, while the last two essays evaluate broader applications with vast social welfare implications. The first essay (chapter 2) of this dissertation, in collaboration with Michele Pezzoni and Fabiana Visentin, exploits a dataset on all applicants to a prestigious Swiss grant to explore a central process in academic life: the application for funds. The results suggest that scientists applying to a grant significantly increase their publicationsâ quality and quantity, learn more, and extend their collaboration network. Beyond the effect of applying, receiving the research funds increases the probability of co-authoring with co-applicants, but it does not have any additional effect on other scientific outcomes. These results justified the title of the chapter since, as it is the case in the Olympics, in research grants, âthe important thing is not to win, it is to take part.â The second essay (chapter 3), also in collaboration with Michele Pezzoni and Fabiana Visentin, uses the same empirical context to explore the determinants of knowledge flows among collaborating scientists. The chapter proposes a novel methodology based on journal references to track knowledge flows among researchers working together. The results suggest that geographical distance does not significantly affect the knowledge flows between team members, but the cognitive distance separating two members does. More specifically, there is an inverted U-curve effect of cognitive distance on the learning among team members: the higher the distance between two scientists in terms of subjects studied, the more they exchange knowledge, up to the point when the distance becomes detrimental because they have too little common ground to communicate. The third essay (chapter 4), in collaboration with Boris Thurm, goes beyond the exploration of the determinants for scientistsâ knowledge production and diffusion to delve into the incentives of all individuals to exchange knowledge. The chapter has two major contributions. First, it acts as a literature review of the empirical evidence on non-financial incentives for knowledge diffusion, such as social recognition, career prospects, and moral considerations. Second, the chapter proposes a simple economic model with heterogeneous agents holding both selfish and moral motives to derive some novel policy implications. The last essay (chapter 5), in collaboration with Dominique Foray, delves into a specific case of knowledge diffusion, the integration of machine learning technologies in healthcare. The analysis suggests that machine learning has the potential for spurring innovation in healthcare but faces several institutional levers. Collecting quantitative data on patents and publications, and qualitative data on hospitals, the results show that machine learning affects healthcare in different ways than older information and communication technologies. The apparition of new business models encourages tech giants to enter the healthcare sector. These patterns have the potential to increase social welfare by reducing externalities in terms of innovation complementarities, but they pose new challenges such as competition policy and human capital formation.

EPFL2020

Chargement

Prior knowledge in Kernel methods

Alexei Pozdnoukhov

Machine Learning is a modern and actively developing field of computer science, devoted to extracting and estimating dependencies from empirical data. It combines such fields as statistics, optimization theory and artificial intelligence. In practical tasks, the general aim of Machine Learning is to construct algorithms able to generalize and predict in previously unseen situations based on some set of examples. Given some finite information, Machine Learning provides ways to exract knowledge, describe, explain and predict from data. Kernel Methods are one of the most successful branches of Machine Learning. They allow applying linear algorithms with well-founded properties such as generalization ability, to non-linear real-life problems. Support Vector Machine is a well-known example of a kernel method, which has found a wide range of applications in data analysis nowadays. In many practical applications, some additional prior knowledge is often available. This can be the knowledge about the data domain, invariant transformations, inner geometrical structures in data, some properties of the underlying process, etc. If used smartly, this information can provide significant improvement to any data processing algorithm. Thus, it is important to develop methods for incorporating prior knowledge into data-dependent models. The main objective of this thesis is to investigate approaches towards learning with kernel methods using prior knowledge. Invariant learning with kernel methods is considered in more details. In the first part of the thesis, kernels are developed which incorporate prior knowledge on invariant transformations. They apply when the desired transformation produce an object around every example, assuming that all points in the given object share the same class. Different types of objects, including hard geometrical objects and distributions are considered. These kernels were then applied for images classification with Support Vector Machines. Next, algorithms which specifically include prior knowledge are considered. An algorithm which linearly classifies distributions by their domain was developed. It is constructed such that it allows to apply kernels to solve non-linear tasks. Thus, it combines the discriminative power of support vector machines and the well-developed framework of generative models. It can be applied to a number of real-life tasks which include data represented as distributions. In the last part of the thesis, the use of unlabelled data as a source of prior knowledge is considered. The technique of modelling the unlabelled data with a graph is taken as a baseline from semi-supervised manifold learning. For classification problems, we use this apporach for building graph models of invariant manifolds. For regression problems, we use unlabelled data to take into account the inner geometry of the input space. To conclude, in this thesis we developed a number of approaches for incorporating some prior knowledge into kernel methods. We proposed invariant kernels for existing algorithms, developed new algorithms and adapted a technique taken from semi-supervised learning for invariant learning. In all these cases, links with related state-of-the-art approaches were investigated. Several illustrative experiments were carried out on real data on optical character recognition, face image classification, brain-computer interfaces, and a number of benchmark and synthetic datasets.

EPFL2006

Chargement

Prior Knowledge in Kernel Methods

Alexei Pozdnoukhov

École Polytechnique Fédérale de Lausanne2006

Chargement

Afficher plus

Prior knowledge in Kernel methods

Alexei Pozdnoukhov

EPFL2006

Prior Knowledge in Kernel Methods

Alexei Pozdnoukhov

École Polytechnique Fédérale de Lausanne2006

EPFL2020