In December, theCOP21conference on climate change underscored how critical it is to understand and manage the impact of climate change on humans and the environment in which we live.

No discussion of that kind would be complete without looking to the ocean, which covers more than 70% of Earth’s surface and holds an estimated 50 – 80% of all life on Earth. [Source] Understanding how the ocean’s food web responds to shifts in temperature, acidity, pressure, etc., is a critical component of understanding the impacts of climate change. To monitor the food chain effectively, we must start at its base, the first link if you will – plankton. These microscopic organisms supply food for larger marine animals, convert sunlight into energy, and so much more, which is why tracking their population dynamics is imperative.

Through the first annual Data Science Bowl, we were given a tool to help us more effectively classify and study plankton species. At the end of the 90 day competition, we emerged with a fundamentally different kind of image classification technique for our field.

为什么这个很重要?在海洋生态,上一页ious methods for analyzing plankton imaging data ranged from time-consuming manual analyses to semi-automated classification. The scientific process was often delayed by months to years as a result, meaning that our insights into ocean health were often delayed, as were our actions. Within the plankton imaging/ecology field, our needs had progressed to a point where we were no longer able to solve them in-house, which is why the timing for the Data Science Bowl was so perfect.

With any one of the top solutions from the Data Science Bowl, we would have gotten significant increases in how accurately we could classify species or types of plankton, which has direct benefits to the speed with which our science can progress. But what is even better is that the Data Science Bowl allows us to connect with the top machine learning and computer vision researchers in the world to continue solving this global problem. Image classification algorithms can then be progressively improved and adapted for a variety of image types, which can increase the applicability of the Data Science Bowl solutions. There is already a cascade of activity in the Plankton Lab at Oregon State University, as well as at the University of Pierre and Marie Curie (France). At both universities, researchers are actively utilizing the solutions from the competition in a variety of plankton imaging systems.

The Data Science Bowl and its participants are playing a substantial part in facilitating the rapid analysis of data from the marine environment. It also illustrated the importance of multi-disciplinary teams working together to take a step towards solving one of our world’s most critical challenges, climate change.

Check out这个视频on the first annual Data Science Bowl.