Thread for Blogs (Just created one for ResNet)


(Kerem Turgutlu) #82

I’ve made some changes regarding to your suggestions. But there may be some mistakes or wrong interpretation. Please let me know when you have time to take a look at it. Thank you so much !


(Pranjal Yadav) #83

Hello everyone,

Here is my contribution in spreading the knowledge.

You can review the 3 most recent posts. Please share your feedback so that I can update posts.

My idea is to create a series of 7 posts covering all possible architectures mentioned in post 1. After completing these 7 posts, I’ll re-visit each of them in the same order, but this time with codes (pytorch, tensorflow) and implementation.

In the meantime, I’m trying to understand pytorch better, so that I can contribute to fastai library development. I would like to request @jeremy to create a dev branch for any feature development. As of now, I am confused between branches and most of the activity occurs on master branch.

Last but not the least, I request you all to suggest possible changes which can better the outcome of this series of posts, which is, “A deeper understanding of Neural Networks”


(Jeremy Howard) #84

Ah OK well your understanding is exactly right. I may have read something into your text that wasn’t there, but I kinda thought you were saying 1-hot encoding had some fundamental deficiency in terms of what it could represent.


(Kerem Turgutlu) #85

Yeah I made some changes maybe that was it, thank you so much for your help. Ok now the post is public https://medium.com/@keremturgutlu/structured-deep-learning-b8ca4138b848 and my twitter handle is @KeremTurgutlu.

Thanks !


(Reshama Shaikh) #86

I’ve created a list of blogs that Jeremy has gone over in the lessons.

If any of these blogs have been written by women, can you let me know? I would like to tweet it out from my Women in Machine Learning & Data Science @wimlds twitter handle. Thanks.

If I’ve forgotten any blogs, or you notice any typos, I would be happy to update.


(Arvind Nagaraj) #87

Thanks @reshama for compiling this!


(sai krishna kalyan) #88

Hello All,
I just wrote my first blog on embedding


I would love to have your feedback.


(Jeremy Howard) #89

@krishnakalyan3 thanks for sharing :slight_smile: A couple of things that could improve this:

  • Use something like Office Lens to redo your photos. It will clean up the contrast and make them much easier to see. Or alternatively, since you’re mainly showing tables, instead you could create the tables in a spreadsheet and format them nicely, and then take a screenshot of that part of the screen. That can look great!
  • Your description makes it sound like a bit like embeddings handle ordinal variables directly. Perhaps it would be helpful to show how embeddings are actually basically doing one-hot encodings “behind the scenes”
  • It would be nice to show examples of how well they work, or what results they create. Check out some pictures from kaggle winner posts or papers, for instance

(sai krishna kalyan) #90

Thanks for the feedback :slight_smile:


(Sanyam Bhutani) #91

@everyone
I have created a weekly-ish newsletter where I’ll be sending the cool DL, CV articles/resources that I stumble across every week. I wanted to ask you guys if it’s okay that I share your articles in that?


(Miguel Perez Michaus) #92

Very instructive, nice to read and interesting!

About the One Hot encoding assumptions, I also read it twice, and I think I found the -possible- issue; From the post, the sentence that begins it all is:

If we one-hot encode or arbitrarily label encode this variable (…)

So two encodings mixed in one sentence. Label encoding does assume equality of distances in ordered levels, or arbitrary distances in nominals. So I would understand this assumption is made by label encoding, not one hot encoding. The way I understand it OHE is just an embedding of dimension one.

Anyway, thanks for the great post, really worth reading! :slightly_smiling_face:


(Kerem Turgutlu) #93

Thanks for catching it! What I tried to emphasize is that with ohe all pair of levels are assumed to be equally similar or dissimilar or distant, where as in label encoding it’s even worse since for example Monday comes just after Sunday but they maybe 7 levels apart. What embeddings allow us is that they learn about level representation vectors so that similar and dissimilarities can be captured in Eucledian space. I might not reflect this sentence correctly I guess. But still thanks for liking it ! I appreciate :slight_smile:


(Kerem Turgutlu) #94

What happens when you accept your medium post to be shared in a channel ? “Towards Data Science” channel invited me to share my article but I don’t know the details of it, should I accept it or not ? Are there any consequences of this situation? Thanks


(Arvind Nagaraj) #95

Accept it…more people will see your post.


(Kerem Turgutlu) #96

Ah ok then, I am overly skeptical with these kind of things :slight_smile:


(Manikanta Yadunanda Sangu) #97

I have added my article to there publication. It went well. But just as a note they will have control over your article along with you. They can make it member only if they wish to. They can edit and do changes to your article. But they didn’t do anything above so I see no issues :slight_smile:


(Sanyam Bhutani) #98

Here is an attempt by me to explain using pretrained networks to predict Sentiment on the IMDB reviews.

Please let me know what do you think about it.
I could really use your guidance if you want to correct me anywhere :slight_smile:

Regards,
Sanyam Bhutani


(Jeremy Howard) #99

You have to know the reputation of the publication. ‘Towards Data Science’ seems quite OK


(Sanyam Bhutani) #100


A short post on Gradient Clipping.


(Vitaly Bushaev) #101

Hey guys! Would appreciate any feedback on last post about gradient descent!