Jeremy AMA

Farah · December 13, 2018, 3:17am

What kind of topics can we expect in part two of the course next spring. How will it be different then part II version 2

rramphal · December 13, 2018, 3:18am

How did you learn to be such a wonderful public speaker? How did you learn to educate in the way that you do?

mkolodny · December 13, 2018, 3:25am

Do you have a favorite architecture for resource-constrained environments?

Would you just use a smaller Resnet? Or is there another architecture you prefer when you need to do inference really quickly?

paul · December 13, 2018, 3:29am

How would you teach deep learning to a 13-year old?

Vishucyrus · December 13, 2018, 3:31am

Hi Jeremy
How to retrain my model after creating it from a different dataset, on a new but few datasets and classes ??
For example, if I created a model with 5 classes and 50 data instances per class. Now I want it to retrain with new data and new class , say 5 data points of 6th class without training it from scratch (i.e without building a whole new databunch with all the data/classes) ??

nikhil_no_1 · December 13, 2018, 3:44am

Hi Jeremy,

What’s your typical day like? How do you manage your time across so many things that you do?

Thank You.

wyquek · December 13, 2018, 4:05am

You mentioned you used to spend 50% of your time learning new things everyday since you were 18. Now you spend 90% of your time doing that. My brain gets toasted after 2 hours of focusing with intensity. Any tips on how you do that for such longer hours?

init_27 · December 13, 2018, 4:11am

Can you tell us about the Super-Bike in your Twitter Cover photo?
(Is that you? :D)

init_27 · December 13, 2018, 4:12am

The library and course are already amazing. What further developments is the fastai team working towards, Can you give us a little teaser?

pnvijay · December 13, 2018, 4:27am

What should be an ideal learning path for a person like me who aspires to be a great DL Practitioner but has a day job with 2-3 hours to invest in learning/practising per day? How can I take what I have learnt in fast.aiV3 Part 1 and take it to the next level?

rohitr · December 13, 2018, 4:34am

How do you see Google’s AutoML and similar services affecting DL practitioners in the future?

cedric · December 13, 2018, 4:37am

Thank you so much for doing an AMA. What are greatest influences on your thinking? Where do you find your inspirations and what are some important principles that guide you in work and in life?

Kaspar · December 13, 2018, 4:58am

when will the internatinal fellowship of part 2 be annonounced

raviteja41 · December 13, 2018, 5:03am

When will be the international fellowship of machine learning course be announced

ecdrid · December 13, 2018, 5:04am

Isn’t DL a more of Plug and play (added tricks, changing few things, substituting one for the other ) and then people try to reason out?
Like isn’t it the reverse of how Science works actually?( First the theory comes)

init_27 · December 13, 2018, 5:09am

If I may speak, @ecdrid I think scientific laws are determined by extensive experimentation as well.
So for both, its experimentation, “Plug and play” driven by intuition/anti-intuitive appraoches.

miwojc · December 13, 2018, 5:17am

part 1 v2 there were many datesets from kaggle, and kaggle was recommended as great way to learn DL.
part 1 v3, only one? dataset form kaggle and i don’t recall hearing recommendation to do kaggle competitions. what’s the reason behind that?

wyquek · December 13, 2018, 5:22am

there is an opposing view, nassim taleb’s book Antifragile would argue that many a times science happens the other way: the practice comes before the theory

jcatanza · December 13, 2018, 5:23am

rpathak · December 13, 2018, 5:25am

How to design custom loss functions.
How to validate if the new loss function is suitable.
How to combine multiple loss functions and validate if they will work well.
After understanding the problem how to say that this loss function will not work but I should combine it with some other loss function