pwc
pwc copied to clipboard
zziz
Comments and Suggestions
Please add your comments and suggestions for improving this repository.
- Which conference/journal papers would you like to see? - What area of Machine Learning is a priority for you? - What area of Machine Learning is missing from this repository? - Do you want to contribute? How can you contribute? - Your general comment/opinion for improving this repository. - Any other form of feedback/comment.
No spam, please.
Thanks for your feedback @EricGzq! What exactly do you have in mind? What kind of categories?
Awesome idea, +1 for open-science.
It would be great to have such a repo for papers in other research communities, such as software engineering. How to proceed? with a new page here?
Thanks for your feedback @monperrus! Sounds promising! Since there are some remaining works to be done here, for now, I have just focused on Machine Learning. After finishing ML, I will move to other research communities. If you want to contribute for say Software Engineering, please do so. Just add a markdown file (such as SoftwareEngineering.md), I will include a link to that in main README file.
I didn't checked all papers, but i looks like their almost all about visual or general ML. An other catogorie then would be NLP.
Thanks for your feedback @Bachstelze! Right now I'm working on the code which I use to maintain this repository. I will include more conferences and journals as soon as I can. NLP, Reinforcement Learning, Speech Processing and other areas of Signal processing and Machine Learning will be included. Plus I will upload all the code and data that is being for this repository.
With respect to the research areas, one way would be to create alternative repositories, such as pwc-se for software engineering.
This would the equivalent of awesome lists referring to awesome-awesome.
Could we separate the papers by year, then by broad categories?
- Image
- Video
- Text
- 3D
Or
- NLP
- Signal Processing
- Algorithms
I realize there is some overlap (Speech-to-text), but at least I could filter through the articles quicker.
Sounds like categorizing is interesting for a lot of people (@monperrus and @Xeoncross also I got some offline comments about it). I will work on it - namely include papers for each category into CATEGORY.md file. The main page will still have all of them together as it does now - I don't want to overcomplicate the main README page.
Dear @MONIKA-KUMAR-JETHANI, you are welcome to work on categorizing part. I have included the raw csv file in /src folder. I will update that file as we go on. But in any case, it can be a starting point for you.
There is https://researchcode.com/ that is similar. Maybe you can put a link to it at the end
Is the data behind https://researchcode.com/ publicly available (as here)?
Dear @furkanusta, thank you for letting me know about it. I have came to know two other similar websites as well (paperswithcode.com and gitxiv.com). I will include all of them in main page soon.
@monperrus I'm not sure, but I don't think so. They might some day, hopefully.
As https://paperswithcode.com , It would be better to mark the DL library for each paper, like TensorFlow, Pytorch, MXNet, ...
Dear @iphysresearch thank you very much for your recommendation. Right now I'm working on collecting papers with code from different conferences. After this stage I will work on categorizing the papers as well as tagging them (which will include implementation framework too).
Hi @zziz. What are your thoughts on building a GitHub pages site? I think it would be easier to implement tags, have an option to search etc in the future. (PS: I could help build one too.)
Dear @raincrash, sounds great! Let's make it! Let's discuss it over an email or conference call.
I can help in building
- GitHub pages site
- Implementation framework tagging on the markdown itself
That is after you take call on what you want to do first. @zziz
Dear @Quantumofcosmos, at this point I'm focused on collecting data basically. Add more papers and codes to the list. This weekend I will discuss website plans with @raincrash. You are welcome to join our discussion (if @raincrash approves of course :) ).
Hi! Sure! Also, @zziz, how are you retrieving the star count? Thoughts on running a cron every week or something, use the GitHub Developer APIs and get the start count updated in the CSV? or are you already doing this?
Hi @raincrash. Yes, I update on Sundays. So tomorrow there will be an update. Yes I do use GitHub API :) Actually I'm going to open source entire system once it's done. Planning to do it till end of the year. Right now it's a bit messy, complicated and many things yet to be implemented :)
I've read through the comments but for emphasis, I will probably repeat some points already made. Here are some additions that I would love to see:
- An additional column for the programming language
- Whether the code is "dockerized"
- Be able to filter/sort the content
- Perhaps some meta-data (broad ML category for instance)
- Adding a link validation script to help identify links that no longer work
Year 2014 , project Tell Me What You See and I will Show You Where It Is , the link to the code seems wrong. Please correct it if it is.
I think BMVC is also a good conference to track? Edit: I noticed there is no ICLR either.
Thanks @zziz for this great repo! I find it very confusing when I want to contribute to this repo. Do I have to update the readme.md or the src folder? I think it's better if you make the pull request more clear, then everyone interested can contribute, which is more efficient.
