Program
Wednesday January 10, 2017
Location: Auditorium 4, MH-Building (MH-bygget), University of Tromsø - The Arctic University of Norway
08:30 Registration opens
09:00 Opening Robert Jenssen
09:15 Keynote 1: Sébastien Lefèvre
10:00 Coffee break
10:40 Modeling Difficulty to Understand Deep Learning Performance, John P Lalor
11:00 Deep Feature Based Discriminative Appearance Model for Multi-Target Tracking, Mohib Ullah
11:20 Possibilities and Challenges of Deep Learning in Automatic Autonomous Vision-based Power Line Inspection, Nhan Van Nguyen
11:40 Lunch
12:40 Keynote 2: Nicolas Courty
13:25 Fine-Tuning of ImageNet trained Deep Neural Networks for Seafloor Mine Recognition in Sonar Images, Narada Dilp Warakagoda
13:45 Panel discussion: Generalization and transferability in Deep Learning
14:30 Coffee break
15:10 Deep Kernelized Autoencoders, Michael Kampffmeyer
15:30 Temporal Kernelized Autoencoders for Time Series with Missing Data, Filippo Maria Bianchi
15:50 Practical information for the northern light activity
18:00 Northern light activity (See info below)
Thursday January 11, 2017
Location: B302, Science-Building (Realfagbygget), University of Tromsø - The Arctic University of Norway
09:00 Keynote 4: Atsuto Maki
10:05 Deep Learning and Radiomics, Cristian C Rocha
10:25 A domain adaptation framework for sequence labeling in Electronic Health Records, Abhyuday Jagannatha
10:45 Coffee break
Location: Auditorium 4, MH-Building (MH-bygget), University of Tromsø - The Arctic University of Norway
11:15 Deep autoencoders for visualizing chronic patient data, Cristina Soguero Ruiz
11:35 Towards deep anchor learning, Mads Adrian Hansen
11:55 Learning to read games, Ulf Brefeld
12:15 Lunch
13:15 Keynote 5: Devis Tuia
14:00 Uncertainty Modelling and Interpretability in Convolutional Neural Networks for Polyp Segmentation, Kristoffer Knutsen Wickstrøm
14:20 Segment-based Credit Scoring using Deep Variational Auto Encoders, Rogelio Andrade Mancisidor
14:40 Coffee break
15:10 Combinations of Deep Neural Nets for Scene Sentiment Analysis, Stephan K Chalup
15:30 Classification of postoperative surgical site infections from blood measurements with missing data using recurrent neural networks, Andreas Storvik Strauman
15:50 Panel discussion: Next steps in Deep Learning
16:30 Leave for Workshop dinner
17:00 Workshop dinner
Northern Lights Social Event
We invite you to our Aurora camp along where our professional guides will host you for the evening. We will have a fantastic vantage point for the Aurora Borealis at our centre, only 25 minutes outside of Tromsø. Not only that but introduce visit to our 300 Huskie and, of course, our puppies.
Relax outside byt the camp fire in and keep warm in the Arctic as we wait for the Northern Lights to dance across the sky.
During the visit we provide a warm meal of Norwegian still Bacalao, local Stock fish (cod) prepared with tomato sauce (vegetarian meal available upon request) served in our traditional round house
Duration: 4 hours
Pickup/Drop off:
Radisson Blu Hotel: 18:45/23:00
Included: Guiding, transport (25 minute each way), dinner, hot drinks, warm clothes;