The 23rd edition of the Geilo Winter School

The lectures focused on a variety of topics related to computational statistics, such as Markov Chain Monte Carlo, Reinforcement learning, Latent Gaussian Models, and Bayesian Methods. Additionally, there were a number of poster presentations from participants, covering a wide range of research topics.

I, an Applied Mathematics PhD student at UiB, presented a poster titled 'Deep Learning for Acoustic Target Classification: Addressing Class-Imbalance with a Similarity-Based Sampling Approach'. My poster discussed how to address the challenge of class imbalance in the sampling of training and validation data for deep convolutional neural networks. The strategy proposed in the poster seeks to equally sample areas containing all different classes while prioritizing background data that have similar characteristics to the foreground class. The Near-Miss algorithm is used to select these tricky areas from the background class in order to detect regions where misclassification is more likely.

At the event, my poster presentation was well-received. Attendees showed great interest in the topic, and I was able to discuss the challenges of class imbalance in deep learning and the proposed algorithm in detail. Additionally, the attendees were intrigued by the CRIMAC project, and I had the chance to talk in more depth about the project and its goals.

The week provided an excellent opportunity to learn and exchange ideas with other participants. In addition to formal lectures and presentations, there was plenty of time for informal conversations and networking. The attendees also had the chance to learn more about the CRIMAC project at the Institute of Marine Research, which I am a part of.

The Geilo Winter School was a great experience and provided an ideal platform to learn more about the latest developments in computational statistics, as well as to network with other young researchers. I am grateful for the opportunity to attend and look forward to the next annual event.