Traffic4cast 2022 competition calls on machine learning community to investigate the complex dynamics of road traffic congestion

Participants were tasked with modeling and predicting traffic congestion in 3 world cities

The Institute of Advanced Research in Artificial Intelligence (IARAI), an independent global machine learning research institute, has announced the opening of its 4e annual Traffic4cast competition in partnership with HERE Technologies. Participants will be challenged to use the latest AI methods to model and predict future traffic congestion levels and vehicle speeds in London, Madrid and Melbourne.

The IARAI Traffic4cast competition is unique in merging AI with real-world datasets and traffic research to advance the understanding of complex traffic dynamics and systems. Winners will receive awards at NeurIPS 2022, the leading conference in AI, and a tome within the Machine learning procedures Research will highlight contributions from the Traffic4cast 2022 Competition.

The competition features two years of real-world data provided by HERE Technologies, derived from billions of fleet GPS points. New this year, participants will use data from traffic loop counters built into the roads of London, Madrid and Melbourne. The goal is to reduce the barriers to using readily available, public loop counter data to predict the future traffic conditions of entire cities.

Traffic4cast 2022

  • Core Challenge – participants are asked to enter the c . predictableonstion level classes (red/yellow/green) for the whole way chart 15 minutes in the future from the past hour using loop counter data only.
  • Extended Challenge – Contestants are invited to the average speeds on each road segment of the chart 15 minutes in the future from the past hour using loop counter data only.

HERE provides attendees with traffic movie clips based on two years of real-world data for London, Madrid and Melbourne. The clips were created using HERE data from more than 100 billion GPS probe points from a large fleet of vehicles. The data is fully anonymized and converted into high-definition movie clips that show traffic over time, frame by frame, including morning, evening, and rush hour traffic events.

The competition brings together researchers around the world from several prominent areas of machine learning, including graph-based modeling, transfer learning, deep learning and time series prediction. Last year’s competition focused on the time domain shift in traffic caused by COVID-19.

Sepp Hochreiter, one of the founders and co-director of IARAI and an artificial intelligence pioneer who invented the neural network architecture for long-short-term memory (LSTM), said: “Building on the three years of success at NeurIPS 2019- 2021, Traffic4cast will continue to improve our understanding of complex traffic systems.This year, researchers of modern machine learning predict traffic congestion in entire cities, simply from the vehicle counters available at selected points.In addition to traffic congestion, advanced models would also predict average speeds on a Our competition will help to advance and leverage the latest techniques in machine learning, such as neural networks in graphs or physics-inspired neural networks.”

“Traffic congestion is a universal challenge that requires in-depth analysis to understand the ‘hidden rules’ that shape vehicle movements. So I’m excited to see what predictive models this year’s participants can generate using this comprehensive new dataset from HERE and the latest advances in AI and machine learning,” said Reinhard Köhn, head of research at HERE Technologies.

The annual Traffic4cast competition receives hundreds of entries from competitive teams around the world, with past winners from South Korea, the US, China and Sweden.

The winners of the core challenge (congestion classes for the whole road graph) will receive the following prizes:

  • 1st place – a voucher or cash prize worth EUR 5,000 for the participant/team plus one free NeurIPS 2022 conference registration.
  • 2nd Place – A voucher or cash prize worth EUR 3,000 for the participant/team plus one free NeurIPS 2022 conference registration.
  • 3rd place – a voucher or cash prize worth EUR 2,000 for the participant/team plus one free NeurIPS 2022 conference registration.

The winners of the extended challenge (average speeds on each road segment) will receive the following prizes:

  • 1st place – a voucher or cash prize worth EUR 5,000 for the participant/team plus one free NeurIPS 2022 conference registration.
  • 2nd Place – A voucher or cash prize worth EUR 3,000 for the participant/team plus one free NeurIPS 2022 conference registration.
  • 3rd place – a voucher or cash prize worth EUR 2,000 for the participant/team plus one free NeurIPS 2022 conference registration.

Entries for this year’s competition must be submitted by October 15, 2022. For more information and participation in Traffic4Cast 2022, click here.

SOURCE: HERE

Leave a Comment

Your email address will not be published.