Achtung:

Sie haben Javascript deaktiviert!
Sie haben versucht eine Funktion zu nutzen, die nur mit Javascript möglich ist. Um sämtliche Funktionalitäten unserer Internetseite zu nutzen, aktivieren Sie bitte Javascript in Ihrem Browser.

CRC 901 – On-The-Fly Computing (OTF Computing) Show image information

CRC 901 – On-The-Fly Computing (OTF Computing)

|

Third SFB 901 Seminar in winter semester 2019/2020

Begin: Wed, 4. of December 2019 (4:00 PM)
Location: Fürstenallee 11, Room F0.530 + F0.225

On December 4, 2019, the 3rd SFB 901 seminar in the winter semester will take place.

4:00 pm - 4:25 pm subproject B2
Speaker: Prof. Dr. Eyke Hüllermeier
Title: On-The-Fly Machine Learning

4:25 pm - 4:50 pm subproject A4
Speaker: Dr. Daniel Kaimann (FG Frick)
Title: Vertical Integration and Digital Platforms

5:00 pm - 5:45 pm parallel Session

Room: F0.530
Speaker: Marcel Wever, subproject B2
Title: On-The-Fly Machine Learning for the Multi-Label Classification Problem

Abstract:
Multi-label classification (MLC) aims at learning a function that is able to assign an instance (commonly represented as a feature vector) a set of relevant class labels, which is a subset of larger set of candidate labels. Due to the complexity of MLC algorithms used for learning such functions from data, choosing a suitable algorithm configuration for a new dataset is a serious challenge -- even for experts. In this talk, we show how our OTF-ML tool ML-Plan for single-label classification tasks, i.e. predicting a single class label for a query instance, can be extended to automate the selection of a suitable algorithm configuration for MLC datasets. Furthermore, we highlight the challenges inherent in this setting and the potential of the On-The-Fly Computing environment to scale parallelisation to provide solutions faster.

Room: F0.225
Speaker: Ilka Tanneberg, subproject A4
Title: The Acceptance and Evaluation of Vertical Integration in Digital Services

Abstract:

In recent years, the number of new digital services for media products has considerably increased. The acceptance of these services depends on the perceived benefits for the customer compared to the already existing services. These benefits can, for example, include the simple transformation of a physical into a digital product for immediate downloadable use or digital business models with on-demand-solutions. The purpose of this research is to analyze digital service models to isolate factors which benefits the customer and increase the acceptance of new digital services. Therefore, we use Amazon review data to identify the correlations between the introduction of various digital services in the media context and the aggregate ratings of customers. The results of the difference-in-differences estimations show that the simple transformation of a physical into a digital good leads to negative valuations while unbundling and subscription-based access positively correlates with the valuation of customers. This suggests that the move from a download to a streaming intermediary strategy leads to positive post-evaluations by customers which, in turn, influence the acceptance of the new digital service.

The University for the Information Society