About

Concept: A message to Researchers & Organization leaders

Many of us became scientists out of wonder for the natural world. Not satisfied by the sheer beauty of the flowers, birds, trees around us, we turned towards the bizarre, the creepy, the unexplained complexity of nature and decided to poke at it. Yet despite this all-encompassing sense of wonder, in the field of biology, most of our time is spent on the type of repetitive, mind-numbing tasks that could make anyone forget why they fell in love with science.

But those time consuming, boring tasks are exactly the type of tasks that current AI could automate: segmentation and labeling of microscope images, classification of sequences (DNA/RNA data, proteins), video data analysis… For now, the dream of a science free of that bloat is still a fantasy. Because biology, due to various factors (including the lack of unified format for data sharing), has not yet fully reaped the benefits of the AI boom. It is on to us to bring this field up to speed, and bring back that sense of wonder to our professional lives.

AI research owes much of its astonishing progress to its culture of open science: the barrier to entry is still much lower than many other fields, with much of the code ready to be run, be tweaked and re-shared for free by anyone with ideas and an internet connection. Various contests and competitions built on open data lead to breakthroughs: the CASP contest gave us AlphaFold, and further back in time, the first public success of deep neural networks was in the ImageNet classification contest.

Other biology-based contests could leverage recent advances in AI for the benefit of biology research, if only we could demonstrate that the benefit of creating unified-format public datasets is worth the immense effort it requires.

That is precisely the goal of The National Institute for Basic Biology‘s Community-Led Science Contest (CoLeSCon). This contest is an event leveraging the public annotated image databases and the wisdom of crowds to generate interest around the fascinating mysteries of biology research. We invite amateurs and professionals alike to share your ideas on how to unearth hidden information from this hundred of millions of pictures and their metadata. What can one do with such a massive amount of data from various fascinating species of plants, animals and fungi all around the world, that cannot be done from one picture alone?

 In the first phase of the context, the best ideas will be selected and awarded prizes ranging from an iPad Pro to paper certificates. In the second phase of the contest, we invite research teams to implement the ideas selected in phase I and present their results, engaging the general public and encouraging independent thinking in the next generation of scientists.

We hope that you share our ideals, that this contest will gain recognition, and that you too will participate and submit your ideas!

Team