commit f53aee000ba7353cc1977d4f5b79f50d09b051ad Author: Mark Claypool <claypool@cs.wpi.edu> Date: Tue May 9 11:17:46 2023 -0400 Updates diff --git a/research-critique.md b/research-critique.md index 6fc052f..8c1330d 100644 --- a/research-critique.md +++ b/research-critique.md @@ -1,6 +1,6 @@ --- pagetitle: Writing a Research Paper Critique -version: 1.3 +version: 1.4 --- # Writing a Research Paper Critique @@ -41,10 +41,10 @@ they are the same. Consider if there is some combination of two different approaches or combinations of results from two different approaches that makes sense - this can often lead to new innovation! -+ **Future work:** List areas of future work that could immediately be -done following the paper. In most papers, the authors themselves list -future work at the end of their paper. Read these - they can also be -in your list! ++ **Identify future work:** List areas of future work that could +immediately be done following the paper. In most papers, the authors +themselves list future work at the end of their paper. Read these - +they can also be in your list! Also, if you find a paper you really like - either in the clarity of the idea and writing, or in the coolness of the approach, or in the commit 0a3a035cc794a63af63c86c0c68decfdb2d808f2 Author: Mark Claypool <claypool@cs.wpi.edu> Date: Mon Aug 8 13:58:58 2022 -0400 Updates diff --git a/research-critique.md b/research-critique.md index 294a7fc..6fc052f 100644 --- a/research-critique.md +++ b/research-critique.md @@ -1,6 +1,9 @@ -# Research Paper Critique +--- +pagetitle: Writing a Research Paper Critique +version: 1.3 +--- -v1.2 +# Writing a Research Paper Critique Reading research papers is a skill. Like most skills, you have the foundation necessary to do the work (how to read, and foundational commit f0e297917cdb8ce64cf688c4ade5035455676b31 Author: Mark Claypool <claypool@cs.wpi.edu> Date: Tue Dec 7 11:51:26 2021 -0500 Updates diff --git a/research-critique.md b/research-critique.md index a05d578..294a7fc 100644 --- a/research-critique.md +++ b/research-critique.md @@ -4,24 +4,25 @@ v1.2 Reading research papers is a skill. Like most skills, you have the foundation necessary to do the work (how to read, and foundational -knowledge of computers and software). What you *can* get better at -is gaining an understanding of and, eventually, improving on the +knowledge of computers and software). What you *can* get better at is +gaining an understanding of and, eventually, improving on the knowledge presented in the papers you read. There is a good, short, paper on how to read a research paper: -> S. Keshav. "How to Read a Paper", ACM SIGCOMM Computer Communication -> Review, Volume 37, Issue 3, Pages 83-84, July 2007. (An updated -> online version is here: <https://tinyurl.com/dgmbwo> +> S. Keshav. "How to Read a Paper", *ACM SIGCOMM Computer +> Communication Review*, Volume 37, Issue 3, Pages 83-84, +> July 2007. (An updated online version is here: +> <https://tinyurl.com/dgmbwo>) For papers I recommend, you should do at least the first two reading passes. For papers you find on your own, you'll do a first pass and, if the work is useful, a second pass. For select papers that are -related to your work, you do the third, in-depth pass. +closely related to your work, you do the third, in-depth pass. For critiquing a paper: -+ **Get clarity:** Make lists of things you don't understand, big or ++ **Get clarity:** Make lists of things you do not understand, big or small. It could be experimental settings, results in graphs or details about the approach. Some could be answered by following on techniques online or related work papers mentioned. Plan on doing @@ -43,9 +44,10 @@ future work at the end of their paper. Read these - they can also be in your list! Also, if you find a paper you really like - either in the clarity of -the idea and writing or in the coolness of the approach or in the -methodology - take particular note. This is often an indication that -this is a good are for you to ground your work. It doesn't mean your +the idea and writing, or in the coolness of the approach, or in the +methodology you could see yourself doing - take particular note. This +is often an indication that this is a good paper for you to read in +depth and base (at least parts of) your work. It doesn't mean your work will be exactly like this paper (but it can!), but it often can -provide guidance as to the kind of work you want to do and the shape +provide guidance as to the kind of work *you* want to do and the shape of the work you will do. commit 13647cf9078d66b1182de2b31c5fbd3f4fbedc0c Author: Mark Claypool <claypool@cs.wpi.edu> Date: Fri Sep 18 09:12:35 2020 -0400 Updates diff --git a/research-critique.md b/research-critique.md index ebbf23a..a05d578 100644 --- a/research-critique.md +++ b/research-critique.md @@ -1,6 +1,6 @@ # Research Paper Critique -v1.1 +v1.2 Reading research papers is a skill. Like most skills, you have the foundation necessary to do the work (how to read, and foundational @@ -12,8 +12,7 @@ There is a good, short, paper on how to read a research paper: > S. Keshav. "How to Read a Paper", ACM SIGCOMM Computer Communication > Review, Volume 37, Issue 3, Pages 83-84, July 2007. (An updated -> online version is here: -> [https://tinyurl.com/dgmbwo](https://tinyurl.com/dgmbwo)) +> online version is here: <https://tinyurl.com/dgmbwo> For papers I recommend, you should do at least the first two reading passes. For papers you find on your own, you'll do a first pass and, commit 1b6f237249e0a07951642a50d0bb522a10207528 Author: Mark Claypool <claypool@cs.wpi.edu> Date: Tue Jan 28 17:14:59 2020 -0500 Updates diff --git a/research-critique.md b/research-critique.md new file mode 100644 index 0000000..ebbf23a --- /dev/null +++ b/research-critique.md @@ -0,0 +1,52 @@ +# Research Paper Critique + +v1.1 + +Reading research papers is a skill. Like most skills, you have the +foundation necessary to do the work (how to read, and foundational +knowledge of computers and software). What you *can* get better at +is gaining an understanding of and, eventually, improving on the +knowledge presented in the papers you read. + +There is a good, short, paper on how to read a research paper: + +> S. Keshav. "How to Read a Paper", ACM SIGCOMM Computer Communication +> Review, Volume 37, Issue 3, Pages 83-84, July 2007. (An updated +> online version is here: +> [https://tinyurl.com/dgmbwo](https://tinyurl.com/dgmbwo)) + +For papers I recommend, you should do at least the first two reading +passes. For papers you find on your own, you'll do a first pass and, +if the work is useful, a second pass. For select papers that are +related to your work, you do the third, in-depth pass. + +For critiquing a paper: + ++ **Get clarity:** Make lists of things you don't understand, big or +small. It could be experimental settings, results in graphs or +details about the approach. Some could be answered by following on +techniques online or related work papers mentioned. Plan on doing +this! + ++ **Criticize the work:** Note explicitly what is wrong with the +approach, what corner cases the authors have not thought of, and/or +what could be done better. This can be ideas both big and small. + ++ **Compare and contrast:** Compare the work with other, similar +papers you have read, noting how the approaches are different and how +they are the same. Consider if there is some combination of two +different approaches or combinations of results from two different +approaches that makes sense - this can often lead to new innovation! + ++ **Future work:** List areas of future work that could immediately be +done following the paper. In most papers, the authors themselves list +future work at the end of their paper. Read these - they can also be +in your list! + +Also, if you find a paper you really like - either in the clarity of +the idea and writing or in the coolness of the approach or in the +methodology - take particular note. This is often an indication that +this is a good are for you to ground your work. It doesn't mean your +work will be exactly like this paper (but it can!), but it often can +provide guidance as to the kind of work you want to do and the shape +of the work you will do.