TheShiftsProject (@ShiftsProject) header
TheShiftsProject (@ShiftsProject) profile picture

We're a community of researchers and ML practitioners focusing on distributional shift. Join our community on Discord: https://xitter.cf/t.co/J3TBacLegs.

Joined July 2022
Photo by TheShiftsProject (@ShiftsProject)Photo by TheShiftsProject (@ShiftsProject)Photo by TheShiftsProject (@ShiftsProject)Photo by TheShiftsProject (@ShiftsProject)
Hi everyone! Come checkout our short tutorial in uncertainty estimation for Trustworthy ML on Dec. 14th 3-6 PM CET. We'll tell you about the Shifts Challenge, the benchmarks and problems to overcome in marine vessel power estimation and MRI imaging. i-aida.org/course/uncertaint…
📢ML people, let's e-talk again: "Uncertainty Estimation for Trustworthy AI" 🗓️ 14 Dec 3PM - 6PM CET 👩‍💻 with @MeriBach, @AndreyMalinin, Vatsal Raina, Nataliia Molchanova, meself! Participants of the @ShiftsProject Challenge 2022: come for Q/A! i-aida.org/course/uncertaint…
1
5
Hi everyone! Come see the tutorial on Uncertainty Estimation for Trustworthy AI and the Q/A sessions at 3PM CET if you can at zhaw.zoom.us/j/68850806290
Hi everyone! Come checkout our short tutorial in uncertainty estimation for Trustworthy ML on Dec. 14th 3-6 PM CET. We'll tell you about the Shifts Challenge, the benchmarks and problems to overcome in marine vessel power estimation and MRI imaging. i-aida.org/course/uncertaint…
1
Hi! We've made two tutorials that take you through all the steps necessary to making a successful submission on GrandChallenge for the Shifts Challenge 2022. If you have any questions - ping us! Vessel Tutorial: tinyurl.com/yc6c8ahr Medical Tutorial: tinyurl.com/56vrhtj2
[1] The Shifts Project is proud to launch the Shifts Challenge 2022, which examines robustness to distributional shift and uncertainty quality in two high-risk tasks of great societal importance. To take part, go to our site on Grand Challenge shifts.grand-challenge.org .
2
11
5
21
[6] Weather and sea conditions that affect vessel power consumption vary based on season, location and vessel state, leading to ubiquitous distributional shift. This can lead to errors in power prediction, fuel planning and route optimisation which can be costly and hazardous.
1
2
4
[7] There is a need for greatly reliability in power consumption modelling. The challenge participants' will be tasked with developing models that are robust to distributional shift and yield accurate uncertainties on vessel data kindly provided by DeepSea.
2
4