The 2019 Symposium of the Norwegian AI Society will take place from May 27 to May 28 at NTNU in Trondheim. The symposium aims at bringing together researchers and practitioners in the field of Artificial Intelligence from Norway and Scandinavia to present on-going work and discuss the future directions of AI. With the symposium NAIS provides a forum for networking among researchers as well building links with related research fields, practitioners and businesses.
|Monday, May 27 2019
|11:00 - 12:00||Registration and Light Lunch
The symposium will take place at Kjelhuset, room Kjel 5, on the Gløshaugen campus.
|12:00 - 12:15||Welcome and Introduction
Massimiliano Ruocco and Kerstin Bach
|12:15 - 13:00||The history and future of the Norwegian AI Society
Agnar Aamodt and Odd Erik Gundersen
|13:00 - 14:00||Technical Session I (chair: Odd Erik Gundersen)
Flávia Dias Casagrande and Evi Zouganeli: Activity Recognition and Prediction in Real Homes
Deepika Verma, Kerstin Bach and Paul Jarle Mork: Similarity Measure Development for Case-Based Reasoning- A Data-driven Approach
Filippo Remonato, Ole Martin Lovvik and Espen Flage-Larsen: Effectiveness of Neural Networks for Research on Novel Thermoelectric Materials. A proof of concept.
|14:00 - 14:45||Invited Talk: Vespa.AI
Abstract: The big data world has mature technologies for offline analysis and learning from data, but have lacked options for making data-driven decisions in real time. This talk introduces vespa.ai - a mature open source platform for storing, selecting processing and making model inferences over large data sets at end user request time.
|14:45 - 15:30||Coffee Break and Poster Session
List of accepted posters
|15:30 - 16:30||Technical Session II (chair: Massimiliano Ruocco)
Tønnes F. Nygaard, Jørgen Nordmoen, Kai Olav Ellefsen, Charles P. Martin, Jim Torresen and Kyrre Glette: Experiences from Real-World Evolution with DyRET: Dynamic Robot for Embodied Testing
Francesco Scibilia, Knut Sebastian Tungland, Anders Røyrøy and Marianne Bryhni Asla: Energy industry perspective on the definition of autonomy for mobile robots
Kai Olav Ellefsen and Jim Torresen: Self-Adapting Goals Allow Transfer of Predictive Models to New Tasks
|16:30 - 17:30||Invited Talk: AI in the renewable energy sector(chair: Odd Erik Gundersen)
Bernt Viggo Matheussen, Agder Energi
Abstract: This talk will focus on trends and challenges for industrial applications of machine learning. It will discuss price forecasting, hydropower optimisation and snow estimation.
|19:00||Symposium Dinner at Bakklandet
|Tuesday, May 28 2019
|9:00-10:00||Panel: The status of AI in Norway: challenges and future directions (chair: Kerstin Bach)
Ole Jakob Mengshoel (NAIL Director / NTNU)
Klas Pettersen (CEO NORA)
Signe Riemer-Sørensen (Research Scientist, Sintef Digital)
Ola Tørudbakken (SVP Rackscale, Graphcore)
|10:00 - 10:30||Coffee Break and Poster Session
List of accepted posters
|10:30 - 11:30||Technical Session III (chair: Kai Olav Ellefsen)
Hege Haavaldsen, Max Aasbø and Frank Lindseth: Autonomous Vehicle Control: End-to-end Learning in Simulated Urban Environments
Signe Riemer-Sørensen, Jie Wu, Halvor Lie, Svein Sævik and Sang-Woo Kim: Data-driven prediction of vortex-induced vibration response of marine risers subjected to three-dimensional current
Fredrik Vatsendvik, Åsmund Brekke and Frank Lindseth: Multimodal 3D Object Detection from Simulated Pretraining
|11:30 - 12:15||Invited Talk: Bayesian preference learning (chair: Lars Hovind)
Arnoldo Frigessi, University of Oslo
Abstract: Preference data occur when users express comparative opinions about a set of items, by rating, ranking, pair comparing, liking or clicking, indicating very incomplete and possibly self-contradictory opinions. The aim is to estimate for each user the individual ranking of the items. This is an important part of a recommender systems. We rapidly review our work on probabilistic approaches to preference learning, including the Mallows and Bradley-Terry models. We illustrate and discuss the use of these methods, comparing to matgrix factorisation approaches. Particularly interesting is the enhancement of diversity. This is joint work with Qinghua Liu, Marta Crispino, Ida Scheel, Valeria Vitelli, Øystein Sørensen, Elja Arjas.
|12:15 - 13:00||Networking Lunch
Technical papers will have a 20 min presentation, please leave 3-5 min for a discussion.