Swipes For Science
Anisha Keshavan
Research Scientist, Child Mind Institute
Affiliate @ UW eScience Institute
Healthy Brain Network
10,000 MRI scans
Why not algorithms?
45 MPH
STOP!
We should work together
Algorithms
- fast
- sometimes stupid
- needs lots of data to learn
Humans
- slow
- sometimes smart
- learn from sparse data
Braindr
Brain Data Review
Keshavan, Anisha, Jason Yeatman, and Ariel Rokem. "Combining citizen science and deep learning to amplify expertise in neuroimaging." bioRxiv (2018): 363382. (under review)
Instructions
Swipe Left
Swipe Right
Win Prizes
Many Ratings
But often, no agreement
Fail
Pass
Need to "down-weight" some raters
| image id | rater 0 | rater 1 |
|---|---|---|
| 1 | 0 | 1 |
| 2 | 1 | - |
| Expert |
|---|
| 1 |
| 0 |
each rater is a feature
each image
is a sample
missing data
Rater importance vs # ratings
Corrected Ratings
Predict image quality
> 100,000 ratings
> 400 users
0.99 AUC = near perfect
Swipes for Science
Thanks eLife Innovation!
see also: