Demos
A text redaction application powered by multiple domain-specific NER models and Pseudonym-Generation model that automatically detects and replaces sensitive entities in German texts ensuring privacy protection.
Live-on-tape demonstration of forecasting filling levels in a complex real-world sewer system.
Demonstration of our multi-task multi-modal model SoilNet for segmenting and classifying soil horizons within soil profile images.
Research and tooling to extract and output structured data out of unstructured documents (PDF) using open source machine learning options. Tailored to student applications.
Demonstration of our classification model which categorizes Presseschau news articles into standardized Manifesto Project categories.