Use deep learning to automatically estimate stories

[Ryan]

Good day,

Are there any tools out there that use deep learning for automatically estimating (suggesting) story points?  It doesn't seem to be too complex — all it needs to do is use image analysis to determine sophistication of attached designs, then parse title and acceptance criteria, use data from previously estimated stories that appear to be similar to estimate the amount of points.

If there's nothing similar out there at the moment, could you please suggest for Atlassian to implement it for Jira?  The amount of data from previously estimated stories they have to train the model on should be enough to get something like 80-90% accuracy for most stories.

[Matt Doar]

Any one can create a suggestion at https://jira.atlassian.com/browse/JRASERVER

My suspicion is that this task is neither as simple nor as accurate as you believe. But don't let that stop you from suggesting it