The Federal Ministry for Economic Affairs sponsors long-running programs to promote the digitalization of the German economy (Measurement of the Digitalisation of the German Economy). They focus particularly on Artificial Intelligence as an enabler of economic growth; besides workshops and publicity initiatives, this also involves conducting in-depth interviews with practitioners in the field.

Every year investigates a different aspect of digitalization, and 2021 is the year of NLP. LangTec has been selected as a prominent example and invited to contribute the practitioner’s perspective in a semi-structured multi-hour interview. The Q&A process ranged over the topics of typical application areas, achievable levels of quality, requirements on IT platforms, modes of operation, cloud computing, user training and of course data protection.

LangTec was pleased to be able to contribute to the understanding of this dynamic market segment on the part of decision-makers in politics. It was a great opportunity to contribute an extensive ground-level view of our core competence, Semantic Text Processing, to interested scientist stakeholders. The results of all interviews will be published as a case study at the end of 2021.

We are delighted to be hosting our two visiting linguistics professors Dr. Bettina Migge (University College Dublin) and Dr. Britta Schneider (Europa-Universität Viadrina Frankfurt (Oder)) at the LangTec office this week. Aim of their research stay with us is the in-depth exchange on applied methods of computational linguistics and text analytics. Dr. Migge and Dr. Schneider jointly chair the working group “Ideologies, Beliefs, Attitudes” in the research network LITHME (Language In The Human-Machine Era). Together with our colleagues they have engaged in a number of extremely insightful and inspiring discussions about LangTec’s projects and possible intersections between academic linguistics and applied language technology in industry.

Not just a case for the police! The professor of journalism Dr. Thomas Hestermann of the Macromedia University of Applied Sciences in Hamburg and his team of researchers investigate how the police handle mentions of the origin of crime suspects and victims in police press releases. Subject to investigation is a corpus of 1.6 million police press releases containing more than 15.7 million sentences published between 2014 to 2018. The distinction of designations of origin for suspects and victims will be a central factor in the results of this research effort and cannot be accomplished robustly by simple keyword search or collocation analysis.

For this reason, LangTec supports the research effort with expertise in deep semantic text analytics. Analyses will be based on techniques permitting the clear discrimination of crime suspects and victims. Thus, while the textual realisations for denoting crime suspects and victims may be the same, deep semantic differences can be detected based on their differing deep structures.

Microservice architectures enable the successful realisation of complex software systems using modular, heterogenous software components.

In this newly commissioned project LangTec supports a leading provider in the augmentation and personalisation of tv broadcasts by digital content. LangTec’s contribution focuses on the provision of expertise in microservice development in Java and Python as well as the functional improvement and extension of the backend infrastructure.

Economics magazine brand eins wants to make accessible to their readers the entire back catalogue of past articles. The brand eins online archive should support search and browsing by topics across the more than 10,000 texts. So far, a manually curated topic structure already exists, but no mapping of texts into that topic structure. It is in the automation of precisely this mapping that LangTec will be supporting with text analytic expertise.

The objective of this project is to map the texts to these topic categories in an n-to-m assignment. In doing so, topic categories as well as their respective association strengths need to be reported for each text. LangTec aims to implement this mapping with machine learning. A central challenge in that will be to adjust the approach such that complete indexing of all texts can be achieved without additional editorial effort – even in the absence of labelled training data.

After months of working from home the whole team were finally able to meet again in person!
To celebrate the occasion, we all met at the fabulous Seehaus Hamburg situated directly on Alster Lake. We enjoyed some amazing sushi and wonderful views onto Alster Lake. It was great to see our colleagues again – and even a little bit of rain in between didn’t harm our good mood 🙂

We are delighted to announce our next amazing addition to the LangTec team. Welcome, Christoper!

Christoper is joining us as a senior machine learning engineer. He has extensive experience in software development, compute infrastructure maintenance and, of course, a strong academic background in machine learning. We really look forward to working with you and wish you a great start here at LangTec!

Great to have you onboard with us. We absolutely look forward to working with you and can’t wait to get you started on your first projects as computational linguist with us. For your start we have a couple of exciting projects lined up for you ranging from working with dependency parsers and analysing syntax trees to medical ontology integration into deep information extraction. Enjoy!

A significant portion in the development in production systems for text and data analytics is dedicated to data engineering: large volumes of data need to be retrieved from different sources, efficiently pre-processed and passed on to the processing components in the right target formats.
We are happy to support a new client from the aviation sector in this aspect with our technical expertise, beginning at the end of March. This mandate will be fulfilled in cooperation with our cooperation partner Ginkgo Analytics, with whom we are in close contact through the Software Alliance Hamburg.

Great news: after a lengthy proposal phase extending over several years, the BMBF compound project Ophthalmo-AI, scheduled over a period of three years, has officially commenced. Ophthalmo-AI focuses on the intelligent, cooperative diagnosis and therapy support in ophthalmology. The project consortium comprises leading representatives from academia and industry: the German Research Center for Artificial Intelligence, Heidelberg Engineering, Saarland University, and the eye clinics Klinik für Augenheilkunde Sulzbach and Augenzentrum am St. Franziskus-Hospital Münster.

LangTecs contribution to the endeavour is dedicated to:

1. the provision of structured information from unstructured historical medical records, as well as their conversion to suitable features for machine learning
2. the development of a GDPR-compliant anonymisation of personal data
3. the technical integration of analysis results into the project framework

We are looking forward to the collaboration with this illustrious consortium!