DEIB team took part in the EUvsVirus Hackathon
May 12th, 2020
Abstract
Barbara Pernici, Mark Carman and a group of Masters students (Dario Scuratti, Virginia Negri, Stefano Agresti) and PhD student (Donya Rooein) participated in the European hackaton EUvsVirus.
The European Commission hosted a pan-European hackathon to connect civil society, innovators, partners and investors across Europe in order to develop innovative solutions for coronavirus-related challenges.
In the context of the EUvsVirus Hackathon, the DEIB team, together with U Geneva, U Zurich and ETHZ developed the “Social Distancing & Masks”project. In this project, the research group want to use images in tweets to support policymakers and epidemiologists to understand the impact of social distance measures and regulations. This information can be very important for making the right decisions at the right time to minimise the effects of the virus. In particular, the aim is to focus on social distancing and the use of masks in public spaces.
The team has developed a pipeline that starts with a high volume of information, ie. tweets selected with minimal tag filtering. The tweets are fed into sophisticated machine learning algorithms for an initial selection of relevant images, which are then exposed to the crowd (volunteer contributors).
By answering some simple questions about the associated images, contributors can confirm the relevance of the image and add details that AI can not capture. This has two goals: on one side, to provide invaluable information of the compliance and effectiveness of measures against COVID, such as social distancing and use of masks, on the other, to provide clean data to further refine the machine learning algorithms.
For further information please see: https://www.youtube.com/watch?time_continue=8&v=jFHGaNzB85M&feature=emb_logo
The European Commission hosted a pan-European hackathon to connect civil society, innovators, partners and investors across Europe in order to develop innovative solutions for coronavirus-related challenges.
In the context of the EUvsVirus Hackathon, the DEIB team, together with U Geneva, U Zurich and ETHZ developed the “Social Distancing & Masks”project. In this project, the research group want to use images in tweets to support policymakers and epidemiologists to understand the impact of social distance measures and regulations. This information can be very important for making the right decisions at the right time to minimise the effects of the virus. In particular, the aim is to focus on social distancing and the use of masks in public spaces.
The team has developed a pipeline that starts with a high volume of information, ie. tweets selected with minimal tag filtering. The tweets are fed into sophisticated machine learning algorithms for an initial selection of relevant images, which are then exposed to the crowd (volunteer contributors).
By answering some simple questions about the associated images, contributors can confirm the relevance of the image and add details that AI can not capture. This has two goals: on one side, to provide invaluable information of the compliance and effectiveness of measures against COVID, such as social distancing and use of masks, on the other, to provide clean data to further refine the machine learning algorithms.
For further information please see: https://www.youtube.com/watch?time_continue=8&v=jFHGaNzB85M&feature=emb_logo
