In this blog post, I want to point you to some resources that can help you find machine learning papers and also keeping up (as far as that is possible) with the fast pace of the area. Note that this should only be guidance on how to start and is by no means an exhaustive list of resources.
When conducting a literature research, there are two standard resources, which probably almost everyone comes across:
- Google Scholar: the Google search engine for literature research and academic works. Check this short tutorial, for example.
- arXiv: De facto standard database to upload preprints and make papers available freely. Used very heavily in machine learning, since progress is very fast and conference schedules can’t keep up with it. Check this short tutorial, for example.
In addition to these very good resources, there are two more very useful resources that are specialized for machine learning topics:
- arXiv Sanity: A wrapper for arXiv which allows you to skim, search, and sort arXiv much more easily. Additional features include searching for similar papers based on TF-IDF features, better abstract overview and a recent hot paper list. Check out the introduction video.
- Papers with code: Resource for Machine learning papers with code. Very useful to find implementations of algorithms. Check this short tutorial, for example. Note that they recently updated the website to make it even better.
When working with machine learning, finding the right data to train your models is always a crucial part of the process. Here are some resources that might be able to help you with this:
- Google data set: Google project to enable finding data sets by simple queries.
- Cloud/Mesh/RGBD datasets: This is a list of point cloud/mesh/RGBD datasets, by Yulan Guo
- CVonline Image Databases: This is a collated list of image and video databases.
How to keep up to date?
I think Twitter and Reddit are very important and useful resources to keep up to date. Check for example these Reddit channels:
On Twitter you will find many interesting discussions, tips and tricks, but also summaries and demos of recent research papers. If you are interested in material to delve into deep learning, there is also many people posting great resources. Some people I highly recommend following on Twitter are:
- David Ha (@hardmaru): Research scientist at Google Brain, who seems to always know the latest advances in machine learning first
- Andrej Karpathy (@karpathy): Head of AI at Tesla, with great side projects and tweets on recent trends
- Yannic Kilcher (@ykilcher): Regularly creates great video explanations on recent and seminal research papers
- Thomas Wolf (@Thom_Wolf): Leading the science team at HuggingFace, open sourcing state of the art natural language processing
- roadrunner01 (@ak92501): Quite anonymous, however always posting the latest research ideas and own trials
- Sebastian Raschka (@rasbt): Author of “Python Machine Learning”, with great open source side projects
- Chip Huyen (@chipro): Changed her job recently, and has not been posting much since then. However, has great posts on machine learning interviews and practical tips.
- Lex Fridman (@lexfridman): Host of a very good (maybe the best?) AI podcast, with guests such as Andrew Ng, Stephen Wolfram and Elon Musk
- Jeremy Howard (@jeremyphoward): Creator of fast.ai, a great resource for learning machine learning
- Adrian Rosebrock (@PyImageSearch): Owner of a website with tons of great computer vision tutorials
- Josh Gordon (@random_forests): Teaching applied deep learning and working at Google
- Aurélien Geron (@aureliengeron): Former PM of YouTube video classification, author of the great book “Hands-on Machine Learning”
- Chris Albon (@chrisalbon): Creator of machine learning flashcards, which he posts regularly
Of course, this is not an exhaustive list, there are many more great people to follow. Feel free to check who else I am following up on my Twitter account.
Also check this video and all the links in the video description for more resources.
I hope you did find some useful information, and maybe even that you’re motivated now to sign up or use Twitter and Reddit. If you have any questions, ideas or comments, feel free to contact me :-)