Using NLP in the Automotive Industry: Approaches & Applications

Hi everyone,

we are very pleased to invite you to our fifth NLP meetup where we will hear two interesting presentations.

This time we are going to have two great speakers from the Automotive Industry explaining NLP approaches and application across the Automotive Value Chain.

=== Agenda ===

18:00 - Welcome reception
18:30 - [Introduction] Agenda and introduction
18:35 - [Presentation, incl. Q&A] How to improve the performance of your industrial chatbot
19:20 - [Break] Networking and refreshment
19:35 - [Presentation, incl. Q&A] Improving processes in the automotive industry using NLP: Experiences within Volkswagen Data:Lab
20:20 - [Networking] Get-together with sponsored food and drinks

=== FIRST Presentation ===

• Maria Sukhareva: BMW AG

• Topic: How to improve the performance of your industrial chatbot

Industrial chatbots demand stable support infrastructure, simplicity of integration into production and compliance with corporal data protection policies. This frequently leads to the fact that the focus is shifted away from the actual performance of the NLP components. Off-the-shelf close-source solutions do not allow to improve the performance in any other than data-driven manner. If a close-source NLU fails to deliver satisfying results, it is impossible to assess the underlying reasons or modify the algorithm. The developer is then left with only one solution: laboriously produce more annotated data.
As the AI integration became an indispensable component of success for many ventures, many non-tech companies create their own departments of AI and NLP specialists that are willing to contribute to the improvement of NLU algorithms. But while choosing a chatbot solution, it has proven to be difficult to find an optimal trade-off between transparency and stability. In this talk, we present our experience with an open-source NLU solution as well as compare it with a close-source NLU and discuss the advantages and limitations of both. We show data-driven as well as algorithm-driven methods for improvement of NLU performance and demonstrate that the most significant improvement was achieved through the combination of both.

• Bio: Maria Sukhareva is a NLP expert with an extensive experience in research and industry covering a wide variety of topics such as dialogue systems, machine translation, argumentation mining etc. Currently working at the NLP Lab of BMW Group, she is responsible for discovery of NLP challenges and integration of NLP solutions into corporate routine and products.

=== SECOND Presentation ===

• Fabienne Braune, Christoph Ringlstetter: Volkswagen Data Lab

• Topic: Improving processes in the automotive industry using NLP: Experiences within Volkswagen Data:Lab

At Volkswagen Data:Lab, the goal of the NLP team is to work along the value chain of the company and improve processes that involve natural language. Injecting state-of-the-art NLP techniques in this value chain comes with many interesting challenges. We give a perspective on our mode of operation as a scientifically rooted team within a large industrial corporation. In addition, we present examples of our 'toolbox' of NLP techniques specifically tailored to use-cases in the automotive domain.

• Bio: Fabienne Braune:
Fabienne is NLP expert at the Data:Lab in Munich where she applies state-of-the-art NLP techniques to various use-cases in the automotive. Before this, Fabienne was a researcher at LMU in Munich (CIS) where she worked mainly on machine translation and cross-lingual transfer learning.

• Bio: Christoph Ringlstetter:
Heading the Natural Language Processing Expert Center at Volkswagen Data:Lab Munich. Before he was working with Gini, Research Associate at Centrum für Informations- und Sprachverarbeitung (CIS) University of Munich and a Postdoctoral Fellow at the Alberta Machine Intelligence Institute (AMII), University of Alberta, Canada.