Our fourth AMA is with the awesome, Dr. Sebastian Raschka.
Dr. Sebastian has gained his PhD from Michigan State University with Thesis title: Uncovering Hidden Patterns of Molecular Recognition and presently he is Assistant Professor of Statistics at the University of Wisconsin-Madison teaching machine learning and deep learning and will be joining Grid.ai as Lead AI educator next month.
He is also a contributor to open source software and author of the bestselling book Python Machine Learning. His research work and courses can be found here.
On Dec 12 at 10am PT, Sebastian will be joined by Sanyam Bhutani for a live interview session talking about his research work, book, ML, Deep Learning and so much more.
Drop in your questions as reply to this thread - you can post anytime from now till the start of the event.
At 10am on Dec 12, Sebastian will answer all questions in the Youtube Livestream.
In your opinion, what are the main skills that a data scientist/ML engineer should have?
The company that I’m working on is starting to create a data science team. We have almost no experience with how to get customers or projects in the ML field, do you have any recommendation?
There is a lot of stuff out there to learn and to do in the ML field (e.g. work in a personal project, write a blog, compete on Kaggle, learn new things and read papers…). That is sometimes is overwhelming . How do you balance your time between those? How do you deal with the FOMO?
Could you recommend any good resource to learn ML theory only with audio (e.g. a podcast)? Or maybe theory videos that can be followed easily only listening to the audio?
If you were a biologist who self learned machine learning, which direction and topics would you focus on today for the best opportunity to advance your career to do work in genomics? Would you try to find an industry job, or work in an academic lab, or invest the years into a PhD (especially curious if you’d avoid this option)?
… joining Grid.ai (the PyTorch Lightning folks) as Lead AI educator next month… hmmm?
Does this mean leaving Univ of Wisc? If so, I understand. Been there, done that, but with mixed feelings.
How do you propose to educate newbies into DL mysteries? Same as academia? I say NOT.
What about… Focus on domain experts, like a PhD researcher in biology …strong in math, but clueless about DL. Form working groups around clear and exciting datasets in those domains. Groom the groups to be self-sufficient but provide the critical mentoring on DL concepts and techniques. Let the group establish its program for competence in that domain. Then, get out of the way! Your success indicator will be arXiv paper submissions. Oh, and have fun…
Well, my question is mainly about applications (e.g. PhD) for students that think their CV is not representitive of their potential.
To explain it further, I am mostly considering students from developing countries and although I know that we have access to almost infinite amount of resources nowadays, but still, the environment heavily affects the decisions we have to make in our path, in other words, sometimes we have to do things all against the odds!
So, my question is that how we can communicate our “story of life” to universities and professors that be fair and also help them to make a more accurate decision?
Maybe this story which is about student from my country could be an example.
To be honest, I personally thought about it a lot but I do not have a working system that helps both professors and students to make decisions better, so I think your experience as an amazing professor could give me a better perspective of what is desired.
Are there any particular datasets you recommend for genes/proteins/molecules that you think is good for learning ML+genomics/biology/biochemistry in order to build up a profile for genomics opportunities?
My question about the domain adaptation versus fine-tuning in transfer learning. Is there any clear distinction when to use domain adaptation and when to use fine-tuning if we have very small dataset to apply classification.
Is it possible for Sebastian to talk about his journey in ML space so far and what he sees himself in the future? Also, what motivates him to work towards his ideas?
Drop in your questions as reply to this thread - you can post anytime from now till the start of the event.
At 10am on Dec 12, Sebastian will answer all questions in the Youtube Livestream.
Your success indicator will be arXiv paper submissions. Oh, and have fun…