1

From Data Imputation to Data Cleaning - Automated Cleaning of Tabular Data Improves Downstream Predictive Performance

We develop and evaluate an application-agnostic ML-based data cleaning approach using well-established imputation techniques for automated detection and cleaning of erroneous values. To improve the degree of automation, we combine imputation techniques with conformal prediction (CP), a model-agnostic and distribution-free method to quantify and calibrate the uncertainty of ML models.

Sebastian Jäger, Felix Bießmann

Automated Extraction of Fine-Grained Standardized Product Information from Unstructured Multilingual Web Data

We implement models that reliably predict product attributes across online shops, languages, or both, and can be used to match product taxonomies between online retailers.

Alexander Flick, Sebastian Jäger, Ivana Trajanovska, Felix Bießmann

GreenDB - A Dataset and Benchmark for Extraction of Sustainability Information of Consumer Goods

We present a second public release of the GreenDB and present a first benchmark for sustainability information extraction.

Sebastian Jäger, Alexander Flick, Jessica Adriana Sanchez Garcia, Kaspar von den Driesch, Karl Brendel, Felix Bießmann

Parallelized Training of Deep NN – Comparison of Current Concepts and Frameworks

Kubernetes based evaluation of TensorFlows’ and MXNet’s throughput, scalability and practical ease of use.

Sebastian Jäger, Hans Peter Zorn, Stefan Igel, Christian Zirpins