Amberscript is going to build an auto summarization model for the VRT, to summarize Dutch new articles into short summaries. There are two approaches in this: extractive, filtering the most relevant sentences from the text, and abstractive, forming entirely new sentences that summarize the text. Amberscript will chose to build the first.
The first step in the creation of such an Extractive Auto Summarization model will be to train an unsupervised extractive summarization model. In the presence of high-quality human-generated abstractive summaries, this approach could be further refined to train an abstractive model using any of the state-of-the-art sequence-to-sequence transformer models.
Usage of Standards for data interoperability:
For the implementation of this, Amberscript makes use of Python, PyTorch and Hugging Face transformers package as well as use pretrained models from Hugging Face and other open source platforms. The models will be fine-tuned on VRT data and other open source datasets using GPUs to allow large-scale and efficient training.
*Expose phase is open to all Experiment phase teams
Involvement in Standardisation Bodies:
ISO27001, ISO9001, GDPR
Timo Behrens holds a Bachelor of Science in electrical engineering from Technical University of Cologne, Germany and Queensland University of Technology, Brisbane. He has experience working as a software developer at startups. He is a technical lead at Amberscript and is leading a team of three engineers. Timo was working on several projects using different technologies and is constantly learning about the latest developments in IT.
Peter-Paul de Leeuw
Peter-Paul holds a MBA from INSEAD and a cum laude Masters degree in Economics & Business from the Erasmus University of Rotterdam. He has experience as a strategy consultant at Amberscript and a consultant at the Organisation for Economic Cooperation and Development (O.E.C.D.) in Paris, France. He is CEO of Amberscript and has led multiple customers to successful projects, with a focus on the educational sector. Peter-Paul holds certifications in Prince2 and Lean Six Sigma green belt and has worked in the Netherlands, China, France. Denmark and the United States. Selected experience: Ministry of Justice – Auto summarization proof of concept University of Utrecht & Amsterdam – Delivering speech recognition and transcription services for qualitative research.
Nithin is a machine learning engineer working on using techniques from NLP and deep learning to improve automatic speech recognition, subtitling, post-processing and auto summarization of transcripts. He holds an MSc in Artificial Intelligence from the University of Amsterdam and is broadly interested in the domain of natural language processing, deep learning and machine learning and continue to invest time in research.
Jolien de Louw
Jolien is the operations lead at Amberscript and holds a BSc & MSc of Economics at Utrecht University, the Netherlands. She has >7 years of experience as a business consultant with Accenture in Europe and the USA.