The AnDi Challenge strikes back!
The 2nd Anomalous Diffusion (AnDi) Challenge aims to evaluate methods for detecting and quantifying changes in single-particle motion. Joins us at:
https://codalab.lisn.upsaclay.fr/competitions/16618
What is new?
- Heterogeneous behavior: The quest to detect changes in single-particle motion is more than a technical challenge: it’s key to unraveling biological function. Your insights could lead to breakthroughs in understanding crucial cellular processes.
- Phenomenological models: While theoretical models of diffusion are foundational, the cellular environment adds layers of complexity. In the 2nd AnDi Challenge, we dive into scenarios where particles dynamically interact with their environment – think trapping, confinement, dimerization, and more. This is your chance to tackle real-world complexity head-on!
- Single-particle videos: Can you characterize and quantify changes of motion directly from single-particle videos? This second edition introduces a new feature besides trajectory analysis and challenges you to think beyond traditional methods of tracking and analysis. Let’s push the boundaries of what’s possible!
Participate and Collaborate
The challenge will be hosted on Codalab (link above). Dive deeper into the challenge details, datasets, metrics, and timeframes by reading our comprehensive manuscript. The manuscript has already passed Stage 1 review and has been in-principle accepted as a Registered Report in Nature Communications.
We are not just challenging your technical skills; we’re inviting you to contribute to a greater understanding of the microscopic world. Whether you are a seasoned researcher or a curious newcomer, your perspective is valuable. Join us in this exciting journey of discovery and innovation. Let’s uncover the secrets of anomalous diffusion and make a lasting impact on the scientific community!
AnDi 2 Seminar
An end-to-end guide to the challenge was broadcasted on February 22nd 2024. We explained how the data looks like (both videos and trajectories), the evaluation metrics and also how to create a proper submission. If you missed it, you can rewatch the video recording of the presentation with this link. If you want to check the slides of the presentation, follow this link.
Challenge Final Leaderboars
The challenge came to an end on June 15th 2024. Here are the final leaderboards for every track and task. You can reorder the columns to know which Team did better at which metric. See the plots at the bottom for more detail!
We will keep updating this section to include more and more info about the teams and results. Stay tuned!
Video track
Single trajectory task
Global Rank Team RMSE (CP) JSC (CP) MAE (alpha) MSLE (K) F1 (diff. type) MRR
1 SU-FIONA 2.948 0.301 0.224 0.578 0.91 0.733
2 ICSO UPV 2.508 0.216 0.425 0.185 0.827 0.6
3 SPT-HIT 2.8 0.272 0.297 0.288 0.848 0.5
Ensemble task
Global Rank Team W1 (alpha): W1 (K): MRR
1 SPT-HIT 0.259 0.058 0.4
2 BIOMED-UCA 0.273 0.33 0.167
2 ICSO UPV 0.38 0.143 0.167
Trajectory track
Single trajectory task
Global Rank Team RMSE (CP) JSC (CP) MAE (alpha) MSLE (K) F1 (diff. type) MRR
1 UCL SAM 1.639 0.703 0.175 0.015 0.968 1.0
2 SPT-HIT 1.693 0.65 0.217 0.022 0.915 0.358
3 HNU 1.658 0.482 0.178 0.06 0.871 0.26
4 M3 1.738 0.649 0.184 0.024 0.652 0.22
5 bjyong 1.896 0.664 0.211 0.252 0.879 0.202
6 SU-FIONA 2.426 0.579 0.194 0.024 0.885 0.199
7 Unfriendly AI 1.709 0.632 0.241 0.042 0.903 0.187
8 KCL 1.803 0.533 0.214 0.03 0.882 0.145
9 EmetBrown 4.439 0.084 0.309 0.056 0.91 0.129
10 BIOMED-UCA 2.138 0.57 0.275 0.026 0.859 0.127
11 Nanoninjas 3.677 0.246 0.203 1.87 0.899 0.126
12 KNU-ON 2.659 0.488 0.307 0.031 0.756 0.101
13 HSC AI 4.025 0.193 0.393 1.402 0.896 0.087
14 AIntgonnawork 1.994 0.447 0.563 0.298 0.545 0.083
15 ICSO UPV 4.056 0.211 0.38 5.255 0.861 0.072
16 D.AnDi 3.879 0.204 0.472 0.275 0.354 0.07
17 DeepSPT 4.894 0.186 0.336 0.424 0.822 0.069
18 far_naz 3.829 0.023 1.476 94.564 0.293 0.06
Ensemble task
Global Rank Team W1 (alpha): W1 (K): MRR
1 UCL SAM 0.138 0.058 0.25
2 DeepSPT 0.267 0.05 0.222
3 Nanoninjas 0.192 0.051 0.167
4 bjyong 0.188 0.084 0.129
5 SPT-HIT 0.231 0.057 0.1
6 ICSO UPV 0.218 0.067 0.09
7 HSC AI 0.23 0.256 0.065
8 Unfriendly AI 0.238 0.072 0.062
9 EmetBrown 0.259 0.622 0.043
10 BIOMED-UCA 0.275 0.534 0.042
11 KCL 0.448 0.593 0.038
Metrics summary
In the plots below you can see the scatter of the various teams that participated in the Challenge over the various metrics (left plot is for Ensemble task, central and right plot are for Single trajectory task, with the colors defining the tracks).