I’ll leave some time free this evening for you folks to ask me about anything at all! Preferably, something I may be competent to answer… Please ask your questions in this thread, and if you see a question you’d like answered, up-vote it. Rachel will ask me the highest-voted questions.
What are the future plans for fast.ai and this course? Will there be a fast.ai Part 3 ?
If there is a part 3 I would really really love to take it.
I’d please like some feedback on the Quora STS task experiments we have been doing. It’s clear that there’s a ton of opportunities in NLP, but some ideas/pointers would really help us focus our efforts in the right areas. Thanks again!
Could you please do a similar course in other countries? like in India/Asia/Europe?
Taken from About Part 2 Category .
To make it a question , I will add “Could you” to the beginning
What was your experience like starting down the path of entrepreneurship? Have you always been an entrepreneur or did you start at a big company and transition to a startup? Did you go from academia to startups or startups to academia?
If you were 25 years old today and still know what you know where would you be looking to use AI?
What are you working on right now or looking to work on in the next 2 years ?
Do you know of any work dealing with 3D objects (like point collections or meshes), instead of 2D images?
It would be interesting to look at applying style transformation to a 2D image (or set) to make it 3D
Would love to hear more about unbalanced classes (stratified sampling, oversampling minority class, tweaking thresholds) or some of other SOTA techniques and hints on how to start modifying fastai classes for this.
Given what you’ve shown us about applying transfer learning from image recognition to NLP, there looks to be a lot of value in paying attention to all of the developments that happen across the whole ML field and that if you were to focus in one area you might miss out on some great advances in other concentrations. How do you stay aware of all of the advancements across the field while still having time to dig in deep to your specific domains?
With all the advancements that we’re able to apply to medical data, how do you go about getting access to additional datasets and working with HIPAA and other regulations?
Very interested to hear more about most promising deep learning methods that take advantage of multi-type datasets.
Would you show us a hands on example of what’s described in https://www.oreilly.com/ideas/drivetrain-approach-data-products? We’ve learned various example of the Modeler - but what would Simulator and Optimizer actually look like?
Can you give us a brief explanation on how would you approach survival analysis. I would like to look into student attrition.
Also, what’s your opinion on data augmentation using gans?
One of my main questions is to gather my data.
I’m interested in tools to OCR, tools to extract, tools to convert.
Is there any way to elaborate on this in the future?
Give some advice about pursuing Machine Learning / Deep Learning PhD. It is worthwhile, what are the requirements, how it compares to working in industry etc?
@jeremy do you envision fast.ai library to become full lifecycle deep learning library some day or would like to keep it focused more as cutting edge training and research tool?
Reinforcement Learning popularity has been on a gradual rise in the recent past. Also quite a lot of people have had differing opinions/debates on Twitter and otherwise.
What’s your take on RL? Would fast.ai consider covering some ground in popular RL techniques in the future?
What Criteria do you use to choose your shirts for the class?
Do you have any suggestions for production tools that can incorporate our deep learning results into a full data pipeline?