Direkt zum InhaltDirekt zur SucheDirekt zur Navigation
▼ Zielgruppen ▼

Humboldt-Universität zu Berlin - Mathematisch-Naturwissenschaftliche Fakultät - Institut für Informatik

Probevortrag Dissertation: Phillipp Schoppmann

Wann 22.10.2020 ab 13:30 (Europe/Berlin / UTC200) iCal
Wo online: Zoom

Am Donnerstag, 22. Oktober um 13:30 Uhr (s.t.) wird Phillipp Schoppmann den
Probevortrag zu seiner Dissertation halten.

Der Titel der Arbeit ist:

  "Secure Computation Protocols for Privacy-Preserving Machine Learning"

Der Vortrag wird per Zoom stattfinden.

Zugangsdaten (nur mit Infromatik-Account)


Secure Computation Protocols for Privacy-Preserving Machine Learning

Machine learning greatly benefits from the availability of large amounts
of training data, both in terms of the number of samples, and the number
of features per sample. However, aggregating more data under centralized
control is not always possible due to security and privacy concerns,
regulation, or competition. Secure computation protocols promise a
solution to this dilemma, allowing multiple parties to train machine
learning models on their joint datasets while provably preserving the
confidentiality of the inputs. In this talk, we will first look at two
prominent applications, linear regression and document classification.
We present secure computation protocols tailored to these high-level
tasks, as well as a general framework for secure linear algebra.
Finally, we take a look at Vector-OLE, a low-level cryptographic
primitive, and propose a novel protocol that helps speed up a wide range
of secure computation tasks, within private machine learning and beyond.