Colloquium


Our colloquium takes place each Tuesday in B6, A1.01, beginning at 13:45. The colloquium is a joint event of Prof. Stuckenschmidt, Prof. Bizer, Prof. Paulheim, Prof. Ponzetto and Prof. Gemulla.

The colloquium allows PhD students to present their work. Talks should be given in English.

Format suggestion for PhD talks:

   20 min results presentation + 20 min discussion


Colloquium@DWS HWS 2018

Date

Presenter

Title


11.09.
Yaser Oulabi

Profiling the Potential of Web Table Data for Extending Knowledge Bases with Long-tail Entities



18.09.
Oliver Lehmberg

Synthesizing N-ary relations from Web tables
Anne Lauscher

Understanding Scitorics: Towards a Holistic Computational Analysis of Scientific Writing
 

25.09.






02.10.
Sascha Krstanovic

Snapshot Ensembles for Time Series Forecasting
Kiril Gashteovski

Exploring Factual Knowledge Potential in Open Information Extraction Triples

09.10.

Amirhossein Kardoost

Semi-supervised and unsupervised video object segmentation



16.10.

Daniel Ruffinelli

Combining Background Knowledge with Knowledge Graph Embeddings
Yanjie Wang

Learning Sparse Tucker Decomposition for Knowledge Graphs


23.10.

Manuel FinkFine-grained Evaluation of Rule- and Embedding-based Systems for Knowledge Graph Completion



30.10.
Anna Primpeli

WDC Training Data Corpus for Large-Scale Product Matching
Fabian Burzlaff

Evaluating Knowledge-driven Architecture Composition

06.11.

Nicolas Heist

Towards Knowledge Graph Construction from Entity Co-occurrence
Jonathan Kobbe

Using Background Knowledge for Argumentative Relation Classification

13.11.
Sven Hertling

DBkWik: One Knowledge Graph from thousands of Wikis



20.11.

Taha Alhersh

On the combination of IMU and optical flow for action recognition
Christian SchreckenbergerTowards Mining of Trapezoidal Timeseries Data Streams

27.11.

Stefan Kain

Fully Unsupervised Anomaly Detection with Neural Networks?
David Ganshorn


04.12.

Alexander Diete

Exploring Fusion in Multimodal Activity Recognition
Christoph Kilian Theil

 ProFET: Predicting Risk of Firms from Event Transcripts