About

About

Artificial Intelligence (AI) is a constitutive element of technological revolution we are witnessing now. In recent years, AI has become the subject of increased interest not only in science and business, but also in governments and public administration. All due to its growing and largely uncontrolled impact on various areas of our individual and social lives.

NASK (Research and Academic Computer Network) as a leading state research institute operating in the domain of the Ministry of Digital Affairs in Poland intensifies its activity in the area of AI in the scientific and practical dimension. The conference "Dimensions of artificial intelligence. NASK - digitalization - Poland" is one of the demonstration of this activity. The conference program refers to the heterogeneity of the meaning of the term artificial intelligence in its practical purposes. The conference is organized for the first time by the Centre of Artificial Intelligence and Data Analysis NASK (AIDA), newly created at NASK.

Agenda

In the wake of the era of the digital civilization, one should bear in mind fundamental principles of democratic societies. Among those, especially important are: equal opportunity to access education and equal opportunity to acquire skills necessary to navigate in the digital world. The role of governmental institutions - like NASK in Poland - is increasingly important in implementing those principles. In my talk I will review three models of the contemporary digital civilization and discuss the impact of AI in them.
As described in a vulgarized manner in a recent paper (in french),
https://www.telecom-paris.fr/algorithmes-biais-discrimination-et-equite, big data are at the core of recent AI developments but one should have in mind that the conditions under which statistical learning tools are valid are particularly strict. In particular the data generating mecanism has to be controlled and checked, which is rarely the case in practice. Outside a precise framework, the solutions proposed by statistical learning procedures may be invalid and/or subject to very strong bias. These biases may themselves create some discrimination and equity problems, a fact which has been emphasized in the last two years by several "scandals". I will review several sources of bias illustrated by some examples and show how in some case it is possible to correct for such biases.
Iris recognition has served the society as a secure and fast way of personal authentication for more than 25 years. During this time, multiple new research challenges have been identified, one of them being the possibly degrading impact of biology-related changes in the human eye, such as those caused by ocular diseases, but also post-mortem decay processes. This lecture presents the experimentation that quantifies the effects these changes may inflict on iris biometrics systems, proposed countermeasures to neutralize them, as well as new applications of iris recognition.
The Halıcıoğlu Data Science Institute (HDSI) is an independent academic unit at
the University of California at San Diego (UCSD) that was inaugurated in 2018.
It is was launched after a generous gift by Taner Halıcıoğlu, a former UCSD alumnae
who was among the first developers at eBay and Facebook.

HDSI has the mission to lay the groundwork for the scientific foundations of the
emerging discipline of Data Science, develop new methods and infrastructure, and
train students, faculty and industrial partners to use data science in ways that
will allow them to solve some of the world’s most pressing problems.

In contrast to competing institutions, HDSI gives equal emphasis to education,
foundational research (including Data Ethics, Privacy and Security), as well as
applications (with explicit ties to industry). As regards education, the
undergraduate major in Data Science has already more than 500 students enrolled;
graduate degrees, both M.S. and Ph.D., are under consideration.
To forecast occupancy, hoteliers combine time-series models with expert knowledge about local drivers for unusual booking surges or drops. Here, we combine a state-of-the-art time-series model with experimental models for such unusual phenomena. Such models are based on natural language processing of local event descriptions. Research shows that statistical NLP models run on properly prepared data can outperform the deep neural network approach.
Thanks to recent advances in technology, amount of available data and new machine learning approaches the state-of-the-art in Natural Language Processing has improved considerably, allowing for new practical applications, both in science and in business. Polish language has only partly participated in this success. During this talk I want to argue that the cause of this situation lies partly in the inherent difficulty of Polish and other Slavic languages and partly in the lesser number of researchers involved in Polish NLP, lack of available data and pre-established evaluation procedures.
Inspired by SemEval, Kaggle, and other scientific evaluation campaigns, we decided to organize PolEval, which was meant to be a platform, where NLP-related tasks are formulated, openly available training data is prepared and systems developed by teams coming from scientific research institutions and from private companies can compete with each other. This way an objective ranking of methods may be created and an exchange of ideas is finally possible between researchers with academic and business background.
During this talk I would like to summarize the last editions of PolEval, the participants, the results and opportunities during the next edition of this competition.
Popular understanding, expectations and fears of AI among the general public are crucial to the development of a holistic public policy approach to the ongoing technological revolution. What do Polish internet users think of AI-driven services, how do they define them and what do they expect of them and finally which ones do they trust or are suspicious of the most?
A premiere presentation of the latest NASK opinion poll and sociological analysis on the social perception of the AI revolution in Poland.
1. Nowadays face biometrics is widely used as a method for automatic human identification. Latest developments in the field of artificial neural networks boosted the accuracy of face recognition algorithms to the level exceeding human perception. However, there is still a place for improvements. Common challenge in face biometrics is to process effectively data in-the-wild, where heads visible in images can have high pose variability. Can these problems be overcome by the use of 3D templates? Topic of the lecture will cover various methods of encoding spatial information about face structure and its application in biometric systems.
2. Voicelab is a Polish producer of software solutions based on artificial intelligence with experienced team of experts to carry out projects. We have created our own complex solution based on Automatic Speech Recognition (ASR) and Natural Language Understanding (NLU). Our Conversational Intelligence platform, active and passive voice biometrics, as well as a solution to train effective intent recognition models create an environment where solutions like automatic sales effectiveness analysis, fraud detection or dialogue modeling, can be build, tested and developed.
Our products are implemented and tested in the market among customers from banking, medical, insurance, consulting, and other sectors.
1. Problem of fast and reliable classification of text documents is essential to many applications where fast assessment of the document contents is part of decision making process. There are many such examples, among them are: automatic spam detection, analysis of the emotional load in the text, detection of paragraphs containing inappropriate of sensitive information or automatic assignment of incoming office mail to proper departments. Both classical and deep learning based approaches to the subject will be discussed. In case of the neural network based methods, the LSTM networks, their advantages, disadvantages and limitations are going to be explained. The process of word embedding in multidimensional real vector spaces will also be touched. Alternatives to the LSTM will be discussed.
It will be shown how the quality of the classification is affected by the choice of the training approach and how the discardment of the beginnings of the documents affect the final results. Finally the approach based on ensemble learning is going to be presented with accompanying results.
2. The amount of medical data collected has been growing exponentially over the past few decades. This growth in data acquisition has not been, unfortunately, paralleled by the same growth rate of statistical learning methods’ development. In my talk, I will give a brief overview of the analytical methods developed by my group and their applications in the medical and public health areas. Specifically, regularization methods applied to the structural brain imaging data as well as signal processing techniques utilized in extracting physical activity information from the raw accelerometry data will be emphasized. A challenging problem in the brain imaging research is a principled incorporation of information from different imaging modalities in regression models. Frequently, data from each modality is analyzed separately using, for instance, dimensionality reduction techniques, which result in a loss of information. We propose a novel regularization method to estimate the association between the brain structure features and a scalar outcome within the generalized linear regression framework and apply it in classification of the HIV+ and HIV- individuals. Quantification of physical activity in a free-living environment is a challenging task. I will summarize our work utilizing data collected from tri-axial wrist-worn accelerometers quantifying sedentary, upright and ambulatory behavior. A number of algorithms extracting features of physical activity and their association with health outcomes will be presented.

About NASK

About NASK

NASK is a national research institute supervised by the Ministry of Digital Affairs. Our key activities are related to ensuring the security of Internet. CERT Polska, operating within the structure of NASK, is tasked with responding to cyber security threats in the network.
NASK carries out the scientific and research & development activities, mainly in the field of security, reliability and efficiency of the ICT networks. Biometric identity verification methods in the security of services hold important place in our institute. NASK is also the Polish national registry of Internet names in the .pl domain.
Educational activities and popularization of the information society idea play crucial role. NASK Academy conducts unique trainings for companies and institutions with particular focus on the ICT security. For years NASK has been carrying out the European Commission's Safer Internet promoting the safe use of new technologies and the Internet among children and youth.
In 2017 the institute became the operator of the National Education Network (OSE) program, which aim is to connect all the schools in Poland to a fast and secure Internet (ose.gov.pl).

Inez Okulska

Inez Okulska

Researcher at NASK National Research Institute. Working on intersections of AI and humanities, especially focused on Natural Language Processing (NLP). Having her PhD in literary translation, she is about to graduate from Master of Science in Robotics and Automation. Active literary critic and translator, among others for Scientific American Journal. From May till November 2014 she was conducting her research as a visiting researcher at Harvard University, Cambridge, MA.

Zamknij

Robert Kroplewski

Robert Kroplewski

Radca prawny, od 1994 r. praktyk konwergencji i specjalista mediów elektronicznych i usług komunikacji społecznej. Specjalizuje się w zagadnieniach nowych mediów, społeczeństwa informacyjnego, interaktywności w komunikacji społecznej, ochronie prywatności i w prawie do informacji, jak również regulacji i deregulacji rynku mediów i telekomunikacji.

Absolwent Wydziału Prawa i Administracji Uniwersytetu Gdańskiego. Ukończył Studia Podyplomowe w Instytucie Własności Intelektualnej Uniwersytetu Jagiellońskiego oraz Studia Podyplomowe Ochrona Konkurencji, Własność Przemysłowa i Prawo komputerowe Uniwersytetu Jagiellońskiego. Arbiter Sądu Arbitrażowego przy Polskiej Izbie Informatyki i Telekomunikacji, a także Sądu Polubownego przy Krajowej Izbie Gospodarczej. Od 2000 r. właściciel Kancelarii Radcy Prawnego Kroplewski.com. W latach 2004-2006 dyrektor Departamentu Prawnego Krajowej Rady Radiofonii i Telewizji. W latach 2011-2016 dyrektor Biura Prawnego Telewizji Polskiej SA. Od 2006 r ekspert ds. mediów i nowych technologii Instytutu Sobieskiego. W latach 2006-2011 ekspert Prezesa UKE, Ministerstwa Kultury i Dziedzictwa Narodowego, Komisji Sejmowej Kultury i Środków Społecznego Przekazu.

Zamknij

Krzysztof Szubert

Krzysztof Szubert

NASK

Former Secretary of State & The Government Plenipotentiary for the Digital Single Market. Visiting Fellow, University of Oxford. Senior Member, St Antony's College, Oxford. Member of the Multistakeholder Advisory Group (MAG) to the Secretary-General, United Nations. Special Advisor FIPRA International, Brussels. Strategic Advisor National Research Institute NASK. Member of the Council of the National Centre for Research and Development (NCBiR).

Zamknij

Wojciech Pawlak

Wojciech Pawlak

NASK

Strategy Advisor to Director of NASK, Director of AIDA. In the past, researcher at the Institute of Philosophy and Sociology of the Polish Academy of Sciences, as well as a manager and expert in the field of demoscopic research using new technologies and Big Data. As the president of TNS OBOP, he introduced telemetry measurement for television audiences in Poland. For many years associated with the electronic media sector, where he served as strategy director (Polsat), program director (TVP2) and president of the station (Telewizja WP - Wirtualna Polska). Currently responsible for the creation and development of the Centre of Artificial Intelligence and Data Analysis NASK

Zamknij

Jacek Kawalec

Jacek Kawalec

VoiceLAB

Entrepreneur, co-creator of Wirtualna Polska, where he was responsible for technology development. He has been actively investing in startups since 2005. Vice President of VoiceLab.AI, where he works on machine learning, mainly on speech recognition technique, artificial intelligence and natural language processing (NLP).

Zamknij

Bohdan Pawłowicz

Bohdan Pawłowicz

NASK

Bohdan Pawłowicz - Adviser to the NASK's Deputy Director of AIDA - Artificial Intelligence and Data Analysis.
For over 20 years associated with the marketing communication market in Poland. Former GM and Member of the Supervisory Board (earlier the Board) of the IAA - International Advertising Association – Polish Chapter. Judge in KER - Advertising Ethics Committee (the third term). Lecturer in marketing communications at Warsaw and Wrocław Universities, Warsaw School of Economics and SWPS.

Zamknij

Jarosław Harezlak

Jarosław Harezlak

Indiana University Bloomington

Dr. Harezlak is a Professor and an Interim Co-chair of the Department of Epidemiology and Biostatistics at the Indiana University School of Public Health-Bloomington, U.S.A. After graduating from Harvard University and 2 years of post-doctoral training at the Harvard School of Public Health, he joined Indiana University where he has been since. He has held a visiting appointment at Johns Hopkins University, Baltimore, U.S.A. and is an adjunct professor at the University of Wroclaw, Poland. His interests span a number of medical areas including mild traumatic brain injury, neurodegenerative diseases and physical activity as well as statistical areas including semiparametric regression, functional data analysis, brain imaging and intensively-collected longitudinal data. Dr Harezlak serves as an Associate Editor for PLoS One and Biostatistics and Epidemiology .

Zamknij

Łukasz Kobyliński

Łukasz Kobyliński

IPI PAN

Assistant professor at the Institute of Computer Science of the Polish Academy of Sciences, where he manages NLP-related projects in the Linguistic Engineering Group, and Chief Science Officer at Sages. Member of the Program Board of Big Data postgraduate studies conducted in cooperation between Warsaw University of Technology and Sages, as well as Data Science in Management postgraduate studies organized by Koźmiński University and Sages. For many years he has been involved in data analysis and machine learning projects, initially in relation to image processing, and now in application to natural language processing. Particularly interested in corpus linguistics, information extraction, as well as effective processing of large data sets.

Zamknij

Konrad Ciecierski

Konrad Ciecierski

NASK

Konrad Andrzej Ciecierski, Ph.D. Ph.D. with distinction in computer science at Warsaw University of Technology. Specializes in application of machine learning in medical sciences, digital signal processing, natural language processing and in broad spectrum of deep learning applications.
Head of Bioinformatics and Machine Recognition department at NASK institute. Employee of Clinic of Neurosurgery of Maria Sklodowska-Curie Memorial Institute of Oncology. Member of neurosurgical team specializing in treatment of Parkinson Disease and other extrapyramidal disorders.

Zamknij

Mariusz Kamola

Mariusz Kamola

NASK

Mariusz Kamola received his Ph.D. degree in automation from Warsaw University of Technology. He has been continuously with NASK and WUT as researcher, involved in projects of application nature. His recent research interests cover natural language processing, complex network analysis, internet of things, and artificial intelligence.

Zamknij

Weronika Gutfeter

Weronika Gutfeter

NASK

Research Assistant in Biometric and Machine Intelligence Laboratory in Research and Academic Computer Network (NASK). PhD Candidate in Informatics at Warsaw University of Technology. She worked in various projects on implementing machine learning methods in biometric systems. Among the projects worth mentioning are: tool for face identification in video forensic analytics, comparing compression methods of biometric templates on smart cards, validation of photography for identity documents and a multimodal access control system. She was an author of few algorithms submitted to the international conferences: one for face identification in images from outdoor surveillance system and one for ear recognition in-the-wild. Her research interests include face recognition methods invariant to pose variability.

Zamknij

Mateusz Trokielewicz

Mateusz Trokielewicz

NASK

In 2012 got his Bachelor's degree in Biomedical Engineering from the Faculty of Mechatronics at the Warsaw University of Technology, and the Master’s degree also in Biomedical Engineering from the Faculty of Electronics and Information Technology at WUT in 2014. From 2013 he is with the Biometrics and Machine Intelligence Laboratory at the Research and Academic Computer Network NASK. In June 2019 he defended with honors his doctoral dissertation entitled „Iris Recognition Methods Resistant to Biological Changes in the Eye” and got a Ph.D. degree in Computer Science from the Faculty of Electronics and Information Technology. Since 2019 an Assistant Professor at NASK, he took the lead of the Biometric Systems Group. His interests include novel methods and applications of iris biometrics, such as post-mortem iris identification, which he investigates as a part of the international research team together with scientists from the University of Notre Dame in USA and from the Medical University of Warsaw. He authored or co-authored more than twenty peer-reviewed publications, including papers accepted for the top international conferences and journals related to biometrics and computer vision

Zamknij

Filip Konopczyński

Filip Konopczyński

NASK

Researcher and analyst at NASK. University of Warsaw graduate (Law and Cultural Anthropology), he carried out projects for the University of Warsaw, National Bank of Poland, Institute for New Economic Thinking and the Kalecki Foundation. Author of articles published by Przekrój, Kultura Współczesna, Visegrad Insight, Gazeta Wyborcza, Newsweek, Res Publica, Kultura Liberalna, Magazyn Kontakt.

Zamknij

Rafał Lange

Rafał Lange

NASK

Sociologist specializing in social research and data analysis and head of the Market and Opinion Research Department at NASK. The John Paul II University of Lublin alumnus (MA, PhD), former head of the Opinia Research Center, National Youth Research Correspondent for Poland in the Directorate of Youth and Sport of Council of Europe in Strasbourg. He carried out research projects for the Council of Europe, Central Statistical Office of Poland, Chancellery of the Prime Minister of Poland, YMCA, Ministry of Foreign Affairs, Polish Ombudsman for Children, British Council and others.

Zamknij

Dimitris Politis

Dimitris Politis

Halicioglu Data Science Institute, University of California San Diego

Dimitris Politis is currently Distinguished Professor at the Department of Mathematics---also affiliated with the Department of Economics—at the University of California at San Diego (UCSD). He is an internationally known academic statistician working on time series, resampling methods, and nonparametric estimation. He has co-authored over 100 journal papers and several books, including the recent textbook ``Time Series: a First Course with Bootstrap Starter’’ (with T. McElroy), Chapman and Hall/CRC Press, 2019.

Zamknij

Patrice Bertail

Patrice Bertail

Université Paris-Nanterre, TéléComParis-Tech

Patrice Bertail is professor of mathematics (specialized in statistics and probability) at the university Paris-Nanterre. He is researcher at Modal-X in Paris-Nanterre and belongs to the chair Big Data and IA at Telecom-ParisTech. He is the author of around 100 publications and a few books on non-parametric statistics. His fields of interest extend from bootstrap and resampling procedures for dependent data, Markov chains, survey samplings to empirical process and statistical learning.
He has done a lot of applications in the field of food risk assessment.

Zamknij

Jacek Leśkow Ph.D.

Jacek Leśkow Ph.D.

Director of NASK

Jacek Leśkow, Ph.D. Eng., Prof. at the Cracow University of Technology, works also at the Institute of Information and Communication Technology, Cracow University of Technology. In 1997 he received his doctorate at the Institute of Mathematics of the Polish Academy of Sciences and his habilitation in 1999 at the Wrocław University of Science and TechnologyDuring several years he worked at the University of California (Santa Barbara, USA) as Director of the Consulting Laboratory in the Department of Statistics, he was also Vice Rector for Scientific Research at the Higher School of Business - National Louis University in Nowy Sącz. He specializes in artificial intelligence and statistical signal processing and has 55 scientific publications and 5 books. In addition, Prof. Jacek Leśkow was a member of the Steering Committee of the Małopolska Regional Operational Programme for 7 years, and a member of the Committee of Advisers to the U.S. Ambassador in Warsaw for 2 years.

Zamknij