andy99

86 ms·

1. 1711.▲
   by andy99 6y ago

   I made a comment on this too, they are not presenting it in the best light because their example has a finite pool of data so using all of it will give the best performance. You can see the active learning curve is steep at first as the tec
2. 1712.▲
   by andy99 6y ago

   I have a suggestion about the first plot you show in the writeup. From what I can see, it is based on a finite pool of data and so it undersells active learning: performance shoots up as AL finds the interesting points, but then the curve f
3. 1713.▲
   by andy99 6y ago

   I'd add that there is a deep connection between active learning and understanding the "domain of expertise" of a model, for example what inputs are ambiguous or low confidence, and which are out of distribution. E.g. BALD is
4. 1714.▲
   by andy99 6y ago

   I have a playbook - there is so much to like about it, in particular that it runs QNX and with a terminal emulator (I'd have to look up the name) you have a very cool terminal based tablet, even including it's own elVIs VI editor
5. 1715.▲
   by andy99 6y ago

   I'm curious to know the answer to this question as well. I have done a fair bit of work working with organizations to identify ML use cases. When we looked at it from a business process perspective, honestly it didn't go very well
6. 1716.▲
   by andy99 6y ago

   This is a point worth emphasizing. So many proposed "AI" use cases insist on trying to make some kind of end-to-end prediction, rather than let ML do something it is well positioned to do (e.g.) search more effectively, and let a
7. 1717.▲
   by andy99 6y ago

   Here is one of mine: https://github.com/rbitr/ivc Version control for ipython notebooks, the goal is to be able to use a Jupyter notebook in a hacky, iterative way, but track your versions and the output they produced
8. 1718.▲
   by andy99 6y ago

   From what I understand reading the course outline, this is an introductory class. Foundational material in a subject is often "completely disconnected" from the reality of the discipline. If I was an engineer taking a first course
9. 1719.▲
   by andy99 6y ago

   > How do public companies deal with unions when the workforce gets to where it no longer really needs to accomplish things? Ask the manufacturing sector
10. 1720.▲
    by andy99 6y ago

    I think a realistic example of what you mean would be looking at getting support from Uber. If you want to something from their checklist, it's simple, but if you have something that doesn't fit nicely into their FAQ, it is madden
11. 1721.▲
    Learning TensorFlow 2.0 from Scratch(github.com)

    2 points by andy99 6y ago|0 comments
12. 1722.▲
    by andy99 6y ago

    A family member is a chef, and for new jobs you have to "stage" [0] where you come and cook for a (unpaid) shift to see if you are a fit, have the skills, etc. To me this is exactly the same thing, and I expect this is very common
13. 1723.▲
    by andy99 6y ago

    Made me think of this that I have used previously: http://suso.suso.org/xulu/Shellshow I had played with writing a pandoc filter that would add formatting escape sequences based on markdown, your tool does a better job
14. 1724.▲
    by andy99 6y ago

    Thanks for sharing this, I like this idea a lot. I had been playing around with using whktml to make image files out of the ChartSS.css charts for situations that require them. Having a web service like this is a cool solution.
15. 1725.▲
    by andy99 6y ago

    Thanks everybody for all the feedback. Lots of things for me to clean up I see - it's great to have people looking at it
16. 1726.▲
    by andy99 6y ago

    Good point about the line chart. There are a few other ways I considered as well: using a clip-path to make an area chart, or using borders of rotated divs for each segment. I had not seen it done with background gradients the way I did it,
17. 1727.▲
    by andy99 6y ago

    Thanks for the feedback - I agree that I need to do some work on how the axes scale and where the ticks go, and I see the bug you are talking about with the axis being greater than the sum. I think it is due to the placement of the tick. Th
18. 1728.▲
    Show HN: Chartss.css – accessible HTML/CSS charts with markdown support(rbitr.github.io)

    265 points by andy99 6y ago|50 comments