During a project examining the use of machine learning techniques for oil spill detection, we encountered several essential questions that we believe deserve the attention of the research community....

A key challenge in contemporary ecology and conservation is the accurate tracking of the spatial distribution of various human impacts, such as fishing. While coastal fisheries in national waters are closely monitored in some countries, existing maps of fishing effort elsewhere are fraught with uncertainty, especially in remote areas and the High Seas. Better understanding of the behavior of the global fishing fleets is required in order to prioritize and enforce fisheries management and conservation measures worldwide. Satellite-based Automatic Information Systems (S-AIS) are now commonly installed on most ocean-going vessels and have been proposed as a novel tool to explore the movements of fishing fleets in near real time. Here we present approaches to identify fishing activity from S-AIS data for three dominant fishing gear types: trawl, longline and purse seine. Using a large dataset containing worldwide fishing vessel tracks from 2011–2015, we developed three methods to detect and map fishing activities: for trawlers we produced a Hidden Markov Model (HMM) using vessel speed as observation variable. For longliners we have designed a Data Mining (DM) approach using an algorithm inspired from studies on animal movement. For purse seiners a multi-layered filtering strategy based on vessel speed and operation time was implemented. Validation against expert-labeled datasets showed average detection accuracies of 83% for trawler and longliner, and 97% for purse seiner. Our study represents the first comprehensive approach to detect and identify potential fishing behavior for three major gear types operating on a global scale. We hope that this work will enable new efforts to assess the spatial and temporal distribution of global fishing effort and make global fisheries activities transparent to ocean scientists, managers and the public.

Data stream research has grown rapidly over the last decade. Two major features distinguish data stream from batch learning: stream data are generated on the fly, possibly in a fast and variable rate; and the underlying data distribution can be non-stationary, leading to a phenomenon known as concept drift. Therefore, most of the research on data stream classification focuses on proposing efficient models that can adapt to concept drifts and maintain a stable performance over time. However, specifically for the classification task, the majority of such methods rely on the instantaneous availability of true labels for all already classified instances. This is a strong assumption that is rarely fulfilled in practical applications. Hence there is a clear need for efficient methods that can detect concept drifts in an unsupervised way. One possibility is the well-known Kolmogorov-Smirnov test, a statistical hypothesis test that checks whether two samples differ. This work has two main contributions. The first one is the Incremental Kolmogorov-Smirnov algorithm that allows performing the Kolmogorov-Smirnov hypothesis test instantly using two samples that change over time, where the change is an insertion and/or removal of an observation. Our algorithm employs a randomized tree and is able to perform the insertion and removal operations in O(log N) with high probability and calculate the Kolmogorov-Smirnov test in O(1), where N is the number of sample observations. This is a significant speed-up compared to the O(N log N) cost of the non-incremental implementation. The second contribution is the use of the Incremental Kolmogorov-Smirnov test to detect concept drifts without true labels. Classification algorithms adapted to use the test rely on a limited portion of those labels just to update the classification model after a concept drift is detected.