This video is a demonstration of Cross Queries Over Streams or CRISIS. CRISIS has been developed by the Institute for Big Data Analytics at Dalhousie University, in collaboration with Defense Research and Development Canada and Lockheed Martin. The goal of CRISIS is to support maritime related decision making, which is accomplished by processing real-time cross-dataset queries over various streaming, heterogeneous and spatially distributed maritime sensors from multiple IoT infrastructures. This work is part of the MIMIR (Mission-relevant Information Management for Integrated Response) project, which seeks to develop ideas and techniques for seamless integration of data streams of the many IoT sensors infrastructure. More detailed information can be found on the paper “Semantic Integration of Real-Time Heterogeneous Data Streams for Ocean-related Decision Making” published at the NATO Science and Technology Board IST-160 Specialists’ Meeting.
Tool to allow users to select and export AIS data from 2012 up to 2016. Users may filter portions of the dataset by selecting a region of the world or the vessel identification(MMSI). Currently only allows data to be exported to csv format.
Link to the tool: https://solr.research.cs.dal.ca/slice/
Tool to allow users to select AIS data based on the fishing activity pattern. The only form is by selecting a region on a map. Currently only allows data to be exported to csv format.
Link to the tool: https://solr.research.cs.dal.ca/fishingobserver/
We have recently published a paper on the use of Machine Learning to annotate trajectories of moving objects. There is an increasing amount of trajectories data becoming available by the tracking of various moving objects: animals, vessels, vehicles, and humans. However, these large collections of movement data lack semantic annotations, since they are typically done by domain experts in a time-consuming activity. We have shown that Machine Learning techniques can be an effective and efficient tool in the process of semantic annotation of trajectories. Specifically, we demonstrate how active learning allows a minimal set of trajectories to be annotated, while preserving good performance measures. We test several active learning strategies with three different trajectories datasets. The paper, “ANALYTiC: An Active Learning System for Trajectory Classification“, co-authored by Amilcar Soares, Research Associate with the Institute, Chiara Renso from the Italian National research Council (CNR) Lab in Pisa, and Stan Matwin, will appear in IEEE Computer Graphics and Applications. The paper describes the ANALYTiC platform, a web-based interactive tool to visually assist the user in the active learning process over trajectory data.
Link for testing the tool: http://bigdata2.research.cs.dal.ca:8084/
A short video about Big Data
French AI STrategy Report https://www.aiforhumanity.fr/pdfs/MissionVillani_Report_ENG-VF.pdf