SentiOne Research Team

World-class developers, linguists, NLP and big data engineers

Introduction

Technology is part of our DNA – which is why we strive to become the global leader in delivering and developing artificial intelligence solutions for marketing and customer service purposes. Since 2017, SentiOne has been consistently extending its R&D department which aims to create proprietary natural language processing technology, linguistic corpuses and new neural network architectures.

Our team is made up of world-class programmers, linguists, NLP engineers and big data specialists. We build natural language understanding technology in association with leading European academic institutions. Using datasets derived from social monitoring, our system can achieve a high degree of accuracy in internet monitoring and conversation processing on any given subject. This allows for the creation of unique chatbots (both text- and voice-based) which successfully automate customer service work.

We place a large emphasis not only on the business side of our solutions, but also on the scientific one. Each year we participate in many articles, publications and presentations on scientific conferences in Poland and abroad. We are also a member of the CLARIN-PL scientific consortium, which builds an infrastructure and resources for the Polish language.

Team

dr Emilia Kacprzak AI research engineer, science advisor

Handles algorithm implementations, leads research and development projects and designing new technologies, as well as creating documentation and reports, and model evaluation. She also keeps an eye on the current state of NLP and machine learning knowledge. Member of the At SentiOne Emilia is also responsible for cooperation with academic institutions in Poland and abroad. Member of the clarin.pl consortium.

Jakub Klimek Research Team Leader

Team leader and research engineer as well as applied mathematics engineer (Gdańsk University of Technology). His responsibilities include designing technologies and NLP algorithms, building applications and integrating NLP/AI solutions into the SentiOne Automate platform. He has over nine years of experience in building NLP systems such as chatbots, information extraction solutions or text generators. Co-author of US patent US9152623B2 “Natural language processing system and method”. In his spare time he plays Ultimate Frisbee.

Michał Lew Head of AI

Master’s degree in Mathematics (University of Gdańsk, RWTH Aachen). Creator of the SentiOne Automate chatbot platform and its NLU engine. Creator of the first deep learning NLU platform in Poland. Author and co-author of leading architectures and training methodologies for neural networks aimed at detecting user intent (such as SOTA on ATIS). Author of the leading Polish NER and sentiment analysis solutions. He has over six years of experience in creating end-to-end AI and NLP solutions. Author and co-author of scientific publications concerning ML and NLP. At SentiOne his responsibilities include leading the company’s AI development and designing new algorithms, particularly NLU ones. In his spare time he raises his two children and reads fantasy novels.

dr Agnieszka Pluwak PhD - Senti Cognitive Services Grant Coordinator, NLU Engine Product Owner

Responsible for planning linguistic and development work on new semantic models, reporting, scientific publications, and promotion and sales of the SCS project. Co-authored grant applications for new SentiOne programs. She has a decade of experience in working with NLP systems, including multi-lingual chatbots, as well as building semantic and syntax models. Doctorate in corpus linguistics from the Institute of Slavic Studies at the Polish Academy of Sciences. Wrote the book “Ramy semantyczne w tekstach umów najmu” as well as numerous articles on processing formal and casual language.

Michał Jędrzej Stańczyk Research engineer

Applied mathematics engineer (Gdańsk University of Technology). Responsible for implementing and evaluating symbolic models and the tools used in their creation and testing, including domain-specific languages. Co-authored the US patent US9152623B2, “Natural language processing system and method”.

Sandra Dobrzyniak Senior Linguistic Specialist

Finished Roman studies with a specialization in translations at the University of Gdańsk. She has worked with the SentiOne R&D team since 2017, where she participated in implementing the French and Italian languages. She is responsible for preparing and annotating language corpora intended for training machine learning algorithms. Her other responsibilities include the semantic analysis of social media posts and verifying semantic models.

Marta Dzielińska Language Research Manager

Responsible for coordinating the work of linguists cooperating with the Research team. Her daily responsibilities include the linguistic analysis of large datasets and preparing semantic model test reporting. From early 2020 she creates and improves NLU resources used in bot projects. At SentiOne since 2013 - before joining the research team she supported the analytics department in creating client satisfaction reports and social media analysis. She finished second-degree English studies with a specialization in translation and has received a degree in psychology, both from the University of Gdańsk. Her interests include pop culture and horror fiction.

Monika Gęgotek Linguistic Specialist

She finished Czech studies with a specialization in linguistics at the University of Wrocław, as well as ethnology at the University of Łódź. She started her career at SentiOne at the analytics team, where she performed qualitative and quantitative user research. As part of the research team she tests semantic and linguistic models and analyses their performance. She also annotates and creates language corpora and is responsible for Polish, English, and Czech translations within the application.

Mateusz Wyszecki Junior Linguistic Specialist

Master’s degree in English studies with a specialization in natural language processing. He’s responsible for annotating language corpora intended for machine learning, as well as creating rule-based language models. He also evaluates the results generated by these models.

Nat Niemczyk Junior Linguistic Specialist

Finished second-degree English studies with a natural language processing specialization at the University of Gdańsk. Responsible for creating linguistic resources, designing methodologies and annotating language corpora, as well as developing and fine-tuning rule-based language models. Moreover, they verify the accuracy of annotations generated by language models.

Research projects

SentiDeepFusion 

SentiDeepFusion: deep learning approach to automation of multilingual information extraction from colloquial texts for conversational interfaces and customer experience analytics.

Project value: 24 357 995,15 PLN (~5,3M EUR)

Funding amount: 18 836 158,78 PLN (~4,1M EUR)

SentiCognitiveServices 

New generation services for marketing automation and social network management based on artificial intelligence.

Project value: 11 699 568,70 PLN

Funding amount: 8 682 464,57 PLN

SentiConverse 

Development and commercialisation of automatic, conversational interfaces based on AI – chatbots, including intent recognition

Project value: 3 000 000,00 PLN

Funding amount: 2 400 000,00 PLN

Research papers

Obuchowski, A., & Lew, M. (2020). Transformer-Capsule Model for Intent Detection (Student Abstract). Proceedings of the AAAI Conference on Artificial Intelligence, 34(10), 13885-13886. https://doi.org/10.1609/aaai.v34i10.7215

Lew M., Pęzik P.: A Sequential Child-Combination Tree-LSTM Network for Sentiment Analysis. In: Z. Vetulani and P. Paroubek (eds.) Proceedings of the 8th Language & Technology Conference: Human Language Technologies as a Challenge for Computer Science and Linguistics, pp. 397–401, Poznań, Poland. Fundacja Uniwersytetu im. Adama Mickiewicza w Poznaniu, 2017.

Pluwak A., Korczynski W., Kisiel-Dorohinicki M.: Adapting a constituency parser to user-generated content in Polish opinion mining. In: Computer Science, vol.17 pp. 23-44, Kraków 2016.

Pluwak A.: Towards The Application of Speech Act Theory to Opinion Mining. In: Cognitive Studies| Études cognitives, Vol. 16 pp. 33-44, 2016.

Koesten L.-M., Kacprzak E., Tennison J.F.A., Simperl E.: The Trials and Tribulations of Working with Structured Data: a Study on Information Seeking Behaviour. In: CHI ’17- Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems, pp. 1277-1289, Denver, 2017.

Kacprzak E., Koesten L.-M., Ibáñez L.-D., Simperl E., Tennison J.: A Query Log Analysis of Dataset Search. In: Cabot, Jordi, de Virgilio, Roberto, Torlone, Riccardo (eds.) Web Engineering, 17th International Conference, ICWE 2017, Rome, Italy, June 5-8, 2017, Proceedings, pp. 429-436, Springer International Publishing, 2017.

Kacprzak E., Koesten L., Ibanez L.-D., Blount T., Tennison J., Simperl E.: Characterising dataset search — An analysis of search logs and data requests. In: Journal of Web Semantics, Vol. 55, pp. 37-55, 2019.

Kacprzak E., Giménez-García J. M., Piscopo A., Koesten L., Ibáñez L.-D., Tennison J., Simperl E.: Making Sense of Numerical Data-Semantic Labelling of Web Tables. In: EKAW 2018: Knowledge Engineering and Knowledge Management, pp. 163-178, 2018

Kacprzak E., Koesten L., Tennison J., Simperl E.: Characterising Dataset Search Queries. In: Companion Proceedings of the The Web Conference 2018, pp. 1485-1488, 2018.

Koesten L., Kacprzak E., Tennison J., Simperl E.: Collaborative Practices with Structured Data: Do Tools Support What Users Need? In: Proceedings of the 2019 CHI Conference on Human Factors in Computing, ACM, 2019.