shan23chen / paper-reading

Just a reading list

How do I go through CS paper:

CS/informatics papers usually hold six main components:

1. title
2. abstract
3. introduction
4. method
5. experiments
6. conclusion

And I will spend time to run through the paper three times:

First round: Run through the title, abstract, conclusion. Take a look at important figures and tables in the Methods and Experiments section. In this way, you can spend more than ten minutes to understand whether the paper is suitable for your research direction.

Second round: After confirming that the paper is worth reading, you can quickly go through the whole paper. You don’t need to know all the details. You need to understand important figures and tables, know what each part is doing, and circle the relevant literature. If you think the article is too difficult, you can read the cited literature.

The third round: what problem to ask, what method to use to solve this problem. How the experiment was done. Close the article and recall what each section is about.

How to edit this ReadMe

（Using semantics-scholar，Because its clean API）

How to use semantics-scholar's API:

(https://img.shields.io/badge/dynamic/json?label=citation&query=citationCount&url=https%3A%2F%2Fapi.semanticscholar.org%2Fgraph%2Fv1%2Fpaper%2F + last chunk of slash in the url + %3Ffields%3DcitationCount) plus sign not included with no space

Cool blog talk about new techs

What is Diffusion // Diffusion Demo with code // Prompting intro // Contrastive learning and multi-modal // Method call //

What are we reading on next?

Proposed by?	Year	Title	Date	citation
Shan	2022	Competence-based Multimodal Curriculum Learning for Medical Report Generation	Sept,06/22

CV

Read?	Year	Title	what do i think	citation
✅	2012	AlexNet	First big W for Deeplearning

Multimodal

Read?	Year	Title	what do i think	citation
✅	2022	CheXzero	Limited refinement work of CLIP
✅	2019	ConVirt	medical domain and inspired CLIP
✅	2020	ViT	Transformer into CV world
✅	2021	CLIP	unsupervised contrastive learning with huge amount data building a richer semantics

Generation

Read?	Year	Title	what do i think	citation
✅	2014	GAN	First GAN

NLP

Read?	Year	Title	what do i think	citation
✅	2017	Transformer	Best for NLP after (MLP、CNN、RNN)
✅	2022	Do Prompt-Based Models Really Understand the Meaning of Their Prompts?	prompt semantics not as how we think

General New stuffs

Read?	Year	Title	what do i think	citation
✅	2021	AlphaFold 2	atomic level 3D protein prediction