A collection of epic (good) papers to be read in ML (and essentially AI) Research by a wanderer.
#100DaysOfMLResearch #RoadToMLResearcher
These are the papers you can use to improve your ML (and AI) knowledge day by day. Obviously you are not forced to complete it in daily fashion or in a very specific order. Feel free to fork them out. Customize it as you want. Most of the papers are in preprint state, they might not be complete so keep it that way. They are randomly collected from internet and are randomly arranged in day wise manner for concreteness. They are (and will be) being iteratively improved according to the state and the flow of learning. Some topics are off course for an undergraduate but some of them are pretty obvious.
The main aim of ml-papers is to learn (or understand) state of the art (as well as naive) research for modern machine learning (and Artificial Intelligence) applications to science, and is curated such that it will be more beneficial for all range of audiences (beginner, amateur, engineer, expert, etc.). The idea is not to sit back and relax, but to learn and evolve to better ourselves towards a better future. Also to fill the gap between industry/academia research and students. Do something one step at a time.
/* Sorry for vague, ambiguous, and biased attributes. Well, it's a model. And don't you know all models are wrong. */
Attributes:
Day No.: {0, 1, 2, 3, ..., 100}
Paper #: 4 digit unique id for each paper (first 3 digit represents day of the paper, and last digit the number of paper in that day)
Paper Title: "Brief title of paper explaining what exactly it is about"
Level: {E (Easy), M (Medium), H (Hard), Ex (Expert)}
Quality: {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10} // 10 being the highest quality
Day Productivity: in % (After reading the paper, how much productive should you feel?)
| Day No. | Paper # | Paper Title | Level | Quality | Productivity |
|---|---|---|---|---|---|
| 0 | 0001 | Bag of Tricks for Efficient Text Classification | * | * | * % of the day |
| 0 | 0002 | The Consciousness Prior | * | * | * % of the day |
| 0 | 0003 | Distilling the Knowledge in a Neural Network | * | * | * % of the day |
- Complete upto Day-100 each having at least 3 papers [COMPLETED]
- Indicate the level of difficulty to each paper [WORKING ON IT]
- Indicate the quality of each paper e.g. high quality, low quality, technical report, student project, etc. [WORKING]
- Increase the number of papers in a day from 3 to max 10 [NOT DONE YET]
- Classify the paper in such a way that each day will have labelled 3 easy, 3 medium, 4 hard papers [NOT DONE YET]
- Include high quality papers where ever possible; at least one in a day [NOT DONE YET]
- Add more points in this list [NOT DONE YET]
The papers are mostly collected from open sources (mainly arXiv.org). If someone finds it restricted, do notify by a pull request. These are for learning purposes, no commercial use of them allowed. They are provided as it is, without any WARRANTY. COPYRIGHT IS OWNED BY RESPECTIVE AUTHORS, PUBLISHERS, INSTITUTES, ACADEMIA, etc.