Evangelos Kanoulas, chair of the group: ‘We develop machine intelligence and augment it with human intelligence to help people, locate, and act on, the information they need.’
The Information Retrieval Lab (IRLab) is part of the Informatics Institute at the University of Amsterdam. As society becomes more and more digital, bringing the right information to the right people at the right time in a fair and transparent way, has become essential for solid decision-making. IRLab’s aim is to create algorithms, develop tools and evaluation methods that enable the next generation of systems to connect people to information.
One line of research focuses on developing interactive algorithms that support collaborations of people and machines to identify information. They do this for example by developing conversational assistants. Another line of research focuses on machine learning for finding and personalizing information. The researchers create data-driven algorithms that help recommender systems and search engines to understand people’s information needs and to act on this understanding, all by interacting with them. But how can you tell whether the algorithms they develop are better than Google or Siri? A third line of research focuses on modern and reliable ways to evaluate the performance of search, recommendation and conversational algorithms.
The applications of the research are ubiquitous as nearly every domain has become information- and data-driven. The lab’s algorithms can be used in many ways: to find documents on the web or in data repositories; to recommend jobs, products or services; or for discovering games, music or movies for entertainment. There are also applications in the medical domain to help medical professionals conduct systematic reviews and in the legal domain to enable lawyers to search for legal court cases.
IRLab is the leading research group in information retrieval in Europe and one of the top ones in the world. The group has a modern artificial intelligence perspective on the problems they try to solve. This makes them unique in the western world. With forty PhD students and post-docs, IRLab is also one of the largest teams within the field of information retrieval.
Over the last ten years IRLab had a continuous and diverse flow of 30 to 40 researchers from over 15 countries. Many PhD students have gone on to win awards, graduated cum laude, and found top positions in industry or academia, both nationally and internationally.
In recent years, the group received a number of grants, from NWO (including the Vidi, Vici, and Pionier programs), KNAW, and EU Horizon 2020 as well as from industrial sources, such as Bloomberg, Criteo, Google and Microsoft.
IRLab has an entrepreneurial spirit and believes in pursuing great science with great impact. The group is engaged in a number of strategic, long-term public-private partnerships with industrial and governmental stakeholders. Most of these collaborations are organized in ICAI labs: AI for Retail (AIR) Lab Amsterdam, Elsevier Discovery Lab and Police Lab AI.
IRLab’s partners include Ahold Delhaize (Albert Heijn, bol.com), Bloomberg, Booking.com, DPG Group, Elsevier, Google, Huawei Europe, Landelijke Politie, Microsoft Research, Sound and Vision, and Wolters Kluwer. With these partners the group works on data-driven methods to understand content, to analyse and predict user behaviour and to make sense of context and information, all in the setting of search engines and recommender systems.
Members of IRLab have also founded start-up companies, such as 904Lab. The most recent entrepreneurial endeavour is the founding of Ellogon AI, a company that analyses EHR data including textual and tabular information, as well as digital biopsies in order to personalize treatment plans for cancer patients.
In the years to come, IRLab will continue to focus on robust machine learning technology for information retrieval that can scale. Democratizing this type of technology is a key mission of the lab. The human aspect will continue to be essential. By forging new breakthroughs in deep learning and machine learning, IRLab wants to contribute to the creation and evaluation of future search and recommender technology in responsible ways.
IRLab positions itself in the AI and Data Science research themes of the Informatics Institute.