I have been selected as a Software Sustainability Institute fellow π
I'm so stoked to finally share this with you!
With the help of this fellowship, I will create content around π€ sustainable machine learning applications in research!
Here's what that means!
We are delighted to announce the 2022 Software Sustainability Institute Fellows! Join us in welcoming our new ambassadors for better research software, including our first International Fellows π https://t.co/9HFJ9FTirB pic.twitter.com/zh8n46BPfZ
β The Software Sustainability Institute (@SoftwareSaved) January 13, 2022
π₯ Let's start this out with the application video!
It's not my usual style, as you know if you follow my Youtube.
It specifically asks for a 6-minute video of a slide-show presentation.
Can you tell I was nervous?
π The Reproducibility Crisis
Machine learning can do amazing feats. BUT time and time again, we see problems reproducing the findings.
Why?
- Randomness of neural networks
- Software versions
- Different hardware
- Proprietary data
- Proprietary labels
π₯ Sustainable Machine Learning Applications
Let's educate researchers not in ML on how to make their awesome experiments using ML reproducible.
The plan from the video: π οΈ Create a workshop for ML applications π₯ Keep going with YouTube βοΈ Write more articles π£ More Talks
π οΈ The Workshop
What does Workshop for Sustainable Machine Learning applications in Research look like?
Open to ideas, here's a start
- fix random seeds
- provide software versions
- use best practices
- interpretable and explainable AI
Maybe PyData will have me?
π₯ Keep Youtube going
My Youtube revolves around
- π€ machine learning
- πΎ data science
- π Python programming
- π careers
And I love talking about making these accessible. Access to open software is at the core of making research reproducible and good.
βοΈWriting
I wrote 60 articles last year.
These two fall squarely into the content that will help you make machine learning work in your research.
If Your Model Doesn't Work, Make it Smaller
π£ Talks
I have presented a bunch of my research, but that's very different to giving a PyData conference talk.
With the help of the amazing mentors for first-timers like myself, I started the "How to Guarantee ..." - series
π€³ Quick plug
Isn't that neat?!
Conclusion
For this Software Sustainability Institute 2022 fellowship I plan to:
- Create a workshop for ML applications
- Keep going with YouTube
- Write more articles
- Give More Talks
I can't wait! π