On Thursday. The defense will be in English. -------- Forwarded Message -------- Subject: [i-prof] Einladung zur Verteidigung meiner Masterarbeit Date: Mon, 20 Aug 2018 13:31:47 +0200 From: Florian Hartmann <florian.hartmann@fu-berlin.de>To: i-profs@inf.fu-berlin.de, i-wimis@inf.fu-berlin.de, i-studi@inf.fu-berlin.de, renee.zentiks@fu-berlin.de
Sehr geehrte Damen und Herren,hiermit lade ich Sie herzlich zur Verteidigung meiner Masterarbeit mit dem Titel “Federated Learning" ein.
Die Verteidigung findet im Rahmen des Mittagsseminars der Arbeitsgruppe Theoretische Informatik am Donnerstag, den 23.08. um 12:00 Uhr s.t. im SR 055 in der Takustraße 9 statt. Die Arbeit wurde von Prof. Dr. Wolfgang Mulzer betreut, Zweitgutachter ist Prof. Dr. Raúl Rojas.
Ein großer Teil der Arbeit wurde bei Mozilla in Mountain View, Kalifornien, verfasst.
Für Fragen dazu stehe ich auch zur Verfügung. Mit freundlichen Grüßen Florian Hartmann Abstract:Over the past few years, machine learning has revolutionized fields such as computer vision, natural language processing, and speech recognition. Much of this success is based on collecting vast amounts of data, often in privacy-invasive ways. Federated Learning is a new subfield of machine learning that allows training models without collecting the data itself. Instead of sharing data, users collaboratively train a model by only sending weight updates to a server. While this better respects privacy and is more flexible in some situations, it does come at a cost. Naively implementing the concept scales poorly when applied to models with millions of parameters. To make Federated Learning feasible, this thesis proposes changes to the optimization process and explains how dedicated compression methods can be employed. With the use of Differential Privacy techniques, it can be ensured that sending weight updates does not leak significant information about individuals. Furthermore, strategies for additionally personalizing models locally are proposed. To empirically evaluate Federated Learning, a large-scale system was implemented for Mozilla Firefox. 360,000 users helped to train and evaluate a model that aims to improve search results in the
Firefox URL bar. _______________________________________________ Automatischer Mailverteiler an Gruppe 'ml-i-prof-mi'. Hinweise dazu siehe Hilfeseite: https://www.mi.fu-berlin.de/w/Tec/AnkuendigungsVerteiler
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature