Study plan after this course?

Part 1 (2018)
1

Hey all. Just wondering whats worked for people in the past in terms of learning after the course is over?

I’m new to this field but have learnt a lot and need to digest this information properly in the next few months (before I forget). My (next 2-3 months) plan is to focus on past and future Kaggle competitions. Just to get better. Interested in reading papers, but not sure how to go about that.

Any advice from past students on whats really helped them improve will be hugely helpful.

THanks!

17 Likes
2

My plan is also kaggle competitions. That and any tutorials that make sense.

I’m also VERY interested in Part 2 of the course. Is there any solid information on that anywhere?

5 Likes
3

I’m going to write more code (either Kaggle if anything interesting comes up, or just do something with public datasets), also write more blog posts.

3 Likes
5

Go hunting.

3 Likes
6

My Plan is to do more Kaggle competitions as well…also learn more about Pytorch (code mode !!)…

Basically, start preparing for part 2 of the course :slight_smile:

2 Likes
7
  1. Kaggle competitions
  2. Watch lessons again
  3. Follow the forums (hoping people continue to post as they work on Kaggle competitions)
2 Likes
8

I’m looking forward to study Deep Reinforcement Learning (teaching computers to play games, robots to walk, etc…), if anyone else is interested please talk to me :grin:

6 Likes
9

Now that we saw the top to bottom approach, I want to climb from the bottom to top. So here’s my plan,

Code:

  1. Logistic Regression in Numpy
  2. 3 layer Neural Network in Numpy
  3. A 2 Conv, 1 Pool and 1 FC layer network in Numpy
  4. Read, understand and blog about momentum, RMSProp, Adam and Residual Networks.
  5. Face Recognition with TensorFlow/Pytorch on the machine I put together lately.
  6. Language modeling with TensorFlow/Pytorch
  7. Try Kaggle Competition
  8. Attend FastAI part 2.
7 Likes
10
  1. Kaggle competitions
  2. Write an article(s) on Medium
  3. Use what I have learnt on real world problems ::sunglasses:
2 Likes
12
  • Revisit the lectures and the forum discussions
  • Apply the knowledge to real-life use cases
  • Blogging
4 Likes
13

I am interested in this too! Though not sure when I will have time to invest here. My plan is to first revisit the lectures once again and strengthen the concepts taught.

3 Likes
14

My plan is along these lines.

  1. Understand Fastai + Pytorch libraries and contribute to them
  2. Work on many toy problems to build an intuition of which, when, how a model works
  3. Work on structured data and see how n/w architectures affect them
  4. Work on Tamil MNIST
  5. Explore reinforcement learning using Fastai library
  6. Work on real life use cases
  7. Blog about all of the above
6 Likes
15

Is that a metaphor? :slight_smile:

16
  1. Watch few lectures again and try to understand concepts deeply.
  2. Try out few Kaggle competitions, Right now I am working on ISRO MOM dataset (a side project) :wink:
  3. Watch Part 2 V1 and ML lectures
  4. Stay active on forums.
  5. Write some blogs (Haven’t worked on any :frowning:)
  6. Apply these techniques in my work place.
  7. Last, Go hunting for a new job :stuck_out_tongue:

Thanks all of you :slight_smile: may the backprop be with you :wink:

3 Likes
17

We have several awesome datascience and deep learning meetups her in Vancouver, B.C. Having a community of learners is awesome - so I recommend joining or creating such a community!

2 Likes
18

I believe long-term study groups can be very beneficial.

5 Likes
19

With regular participation in Kaggle…

20

Yeah, We all can help each other, study group can be beneficial for us.

21

My plan is to grab some beer…

14 Likes
22
  1. Learn Pandas and scikit
  2. Complete my course todo list and my own notes
  3. Revisit Ng’s DL course
  4. Get more intuition about the techniques used
  5. Participate in a currently running Kaggle competition
  6. deeplearningbook and arxiv papers
1 Like