How to read scientic papers

Introduction - Pervasive Computing Course 2011 IT University of Copenhagen - Aurlien Tabard

mercredi 31 aot 2011-w

Hi, I am Aurlien Tabard Im a post-doc in the SDG group working with Jakob and Morten here. And I spend quite some time in the pitlab. Ill give you a short intro to something you may already be familiar with. But its always good to start the year with some clear outlines. Technical writing is like programming, you need to read a lot to get better at reading. There is no magic potion. But reading it a good way to see how people are doing. For the paper youll have to write at the end of the project, you should really get a paper you like and use it as a model.

How to read scientic papers

Why so many papers? Overview of paper reading. How to go through your reading assignments? 1st, 2nd and 3rd pass.

mercredi 31 aot 2011-w

Why so many papers?

Get you to know and understand the current state of the art in the Pervasive and Ubiquitous computing eld: .: cover more than 20 years of research and exciting technology .: present current questions within the eld .: create interesting new projects

mercredi 31 aot 2011-w

How to Read a Paper

David R. Cheriton School of Computer Science, University of Waterloo Waterloo, ON, Canada

S. Keshav

keshav@uwaterloo.ca

ABSTRACT
Researchers spend a great deal of time reading research papers. However, this skill is rarely taught, leading to much wasted eort. This article outlines a practical and ecient three-pass method for reading research papers. I also describe how to use this method to do a literature survey. Categories and Subject Descriptors: A.1 [Introductory and Survey] General Terms: Documentation. Keywords: Paper, Reading, Hints.

4. Glance over the references, mentally ticking o the ones youve already read At the end of the rst pass, you should be able to answer the ve Cs: 1. Category: What type of paper is this? A measurement paper? An analysis of an existing system? A description of a research prototype? 2. Context: Which other papers is it related to? Which theoretical bases were used to analyze the problem? 3. Correctness: Do the assumptions appear to be valid? 4. Contributions: What are the papers main contributions? 5. Clarity: Is the paper well written? Using this information, you may choose not to read further. This could be because the paper doesnt interest you, or you dont know enough about the area to understand the paper, or that the authors make invalid assumptions. The rst pass is adequate for papers that arent in your research area, but may someday prove relevant. Incidentally, when you write a paper, you can expect most reviewers (and readers) to make only one pass over it. Take care to choose coherent section and sub-section titles and to write concise and comprehensive abstracts. If a reviewer cannot understand the gist after one pass, the paper will likely be rejected; if a reader cannot understand the highlights of the paper after ve minutes, the paper will likely never be read.

1.

INTRODUCTION

Researchers must read papers for several reasons: to review them for a conference or a class, to keep current in their eld, or for a literature survey of a new eld. A typical researcher will likely spend hundreds of hours every year reading papers. Learning to eciently read a paper is a critical but rarely taught skill. Beginning graduate students, therefore, must learn on their own using trial and error. Students waste much eort in the process and are frequently driven to frustration. For many years I have used a simple approach to eciently read papers. This paper describes the three-pass approach and its use in doing a literature survey.

2.

THE THREE-PASS APPROACH

The key idea is that you should read the paper in up to three passes, instead of starting at the beginning and plowing your way to the end. Each pass accomplishes specic goals and builds upon the previous pass: The f irst pass gives you a general idea about the paper. The second pass lets you grasp the papers content, but not its details. The third pass helps you understand the paper in depth.

2.2

The second pass

2.1

The rst pass

In the second pass, read the paper with greater care, but ignore details such as proofs. It helps to jot down the key points, or to make comments in the margins, as you read. 1. Look carefully at the gures, diagrams and other illustrations in the paper. Pay special attention to graphs. Are the axes properly labeled? Are results shown with error bars, so that conclusions are statistically signicant? Common mistakes like these will separate rushed, shoddy work from the truly excellent. 2. Remember to mark relevant unread references for further reading (this is a good way to learn more about the background of the paper).

The rst pass is a quick scan to get a birds-eye view of the paper. You can also decide whether you need to do any more passes. This pass should take about ve to ten minutes and consists of the following steps: 1. Carefully read the title, abstract, and introduction 2. Read the section and sub-section headings, but ignore everything else 3. Read the conclusions

mercredi 31 aot 2011-w

For instance before preparing this I read a bit :)

He are two references you can go to if you want to know more about reading. the rst reference, How to read a paper is linked on the blog, and its only 2 pages. the second is a book from a French Professor of La Sorbonne who explains how to talk about books you havent read. Which he does regularly. But that guy has a trick: he already read a lot and knows a lot about litterature...

Disclaimer

You need to know the domain rst!

...hence read

mercredi 31 aot 2011-w

If you didnt read a specic piece of Shakespeare but read 5 of them and essays about Shakespeare writings then you have some idea about shakespeares plays, the plot, the style, when it came up in his writing and why he was addressing this subject. Why read?

because you have to build your own Universal Library so that you can situate one paper within it. Umberto Eco call the universal library the library containing all knowledge.

How to read scientic papers

Navigate between:

Skimming

Careful reading

mercredi 31 aot 2011-w

How to read scientic papers

Idea of the subject

Good understanding

pass 1

pass 2

pass 3

mercredi 31 aot 2011-w

Active reading

.: Get rid of distractions .: Get a pen .: Jump around, re-read, go backwards - These are not novels .: Talk to others

mercredi 31 aot 2011-w

Reading requires attention, skipping some parts may hinter the overall understanding Reading is active, get a pen, Go in whatever order you want. Authors will (or should) tell you who is the killer from the beginning. Reading is a social activity, we share interesting papers we read, talk to people for advices on papers...

Papers structures

Textbook Vision / Overview Technical paper

mercredi 31 aot 2011-w

Before skimming or reading in details, its good to know that scientic paper share often a very similar structure, that helps readers to browse articles. Textbook:

- Stable knowledge, a few years old. vs. Current/ongoing

- overview vs. Precise (data, experimental process) TODO: SHOW A REAL PAPER SCREENSHOT

GridOrbit An Infrastructure Awareness System for Increasing Contribution in Volunteer Computing

Juan David Hincapi -Ramos, Aur lien Tabard, Jakob E. Bardram e e IT University of Copenhagen DK-2300 Copenhagen, Denmark {jdhr,auta,bardram}@itu.dk
ABSTRACT

The success of a volunteer computing infrastructure depends on the contributions of its users. An example of such an infrastructure is the Mini-Grid, a local peer-to-peer system used for computational analysis of DNA. The speed of analysis increases as more users join the Mini-Grid. However, the invisible nature of such an infrastructure hinders adoption, as it is difcult for users to participate in an infrastructure they are not aware of. This paper introduces GridOrbit, a system designed to increase user awareness, fostering contributions to this infrastructure. We designed GridOrbit using a participatory design process with biologists, and subsequently deployed it for use in a biology laboratory. Our results indicate that the number of contributors to the MiniGrid increased with the use of awareness technologies. In addition, our analysis presents their motives and behaviors. Finally, a characterization of user interaction with GridOrbit emerged, which enabled us to understand how awareness systems can be better designed. We see GridOrbit as an example of a broader class of technologies designed to create Infrastructure Awareness as a means to increase the contributions to technological infrastructures.
Author Keywords

which seek to gain computational power by enlisting enduser computers like PCs and game consoles. Volunteer computing infrastructures however, like other infrastructures, are often invisible to their end-users [16, 22], and this invisibility poses a fundamental obstacle to their adoption [5]. A core question is how end-users can become aware of such an invisible infrastructure, and start participating and contributing to it. Existing volunteer computing projects rely on individual and social motivations, leading to efforts in building communities, setting up competitions, and rewarding users who participate [18]. Thus, there is a substantial overhead associated with recruiting users who will donate CPU cycles for free, and this in turn becomes a core challenge for the scientists using the infrastructure. In this paper, we explore the use of awareness technologies to recruit contributors to a volunteer computing infrastructure in a molecular biology research laboratory. Molecular biologists use the infrastructure to execute bioinformatics algorithms for analyzing DNA/RNA sequences of millions of bytes. This infrastructure uses peer-to-peer (P2P) technology for distributing tasks to computers within the organization. This implies that the infrastructure requires many users to participate. Thus, one central challenge is to motivate users to contribute despite the fact that only a minor part of them have the actual need of submitting tasks. To facilitate recruitment, we aimed at increasing the visibility of the P2P grid. We engaged in a user-centered design process with biologists, which resulted in the design of a system named GridOrbit. GridOrbit displays an interactive visualization of the underlying activity in the infrastructure on public screens (see gure 3), and provides users with feedback about their contributions through notications on their personal computers (see gure 4). GridOrbit was designed with the hypothesis that increasing awareness of a local resource sharing infrastructure will lead to broader participation and increased contribution. We tested this hypothesis in a one-month deployment of the system. Results show that public display visualizations and personal notications can be used for creating an awareness of an otherwise invisible infrastructure. We further show that awareness of the activity of an infrastructure through both public and personal displays supports the recruitment of new contributors. Based on this case, we expand the notion of Infrastructure Awareness and discuss how it can improve the adoption and thereby the value of voluntary infrastructures.

.: Title .: Authors .: Abstract .: Intro .: Related work .: System description .: Evaluation .: Results .: Discussion .: Conclusion .: References

Volunteer Computing, Infrastructures, Infrastructure Awareness, Public Displays, Ambient Displays.

ACM Classication Keywords

H.5.3 Collaborative computing

General Terms

Experimentation
INTRODUCTION

Volunteer computing is a powerful way to conduct computation-intensive data processing by harnessing computing resources from large numbers of geographically distributed individuals. The most prominent examples of volunteer computing initiatives are SETI@Home and Folding@Home1 ,
1

http://setiathome.ssl.berkeley.edu/ http://folding.stanford.edu/

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for prot or commercial advantage and that copies bear this notice and the full citation on the rst page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specic permission and/or a fee. CHI 2011, May 712, 2011, Vancouver, BC, Canada. Copyright 2011 ACM 978-1-4503-0267-8/11/05...$10.00.

mercredi 31 aot 2011-w

10

Questions to keep in mind

.: What questions does the paper address? .: Is the problem relevant? .: What are the main conclusions of the paper? .: What evidence supports those conclusions? .: Do the data actually support the conclusions? .: What is the quality of the evidence? .: Why are the conclusions important?

mercredi 31 aot 2011-w

11

Pass 1 - Overview

Get a general idea Situate the article within the Universal Library.

mercredi 31 aot 2011-w

12

Pass 1 - Overview
Title + Authors .: What is this about + Where does it come from Abstract .: What was done, what is the contribution Sub-sections titles, gures .: What is the paper structure / contribution Medium (textbook, major journal/conference, workshop) .: Who is the audience? References .: Is it a serious paper

mercredi 31 aot 2011-w

13

This can be done almost out of the digital library or google scholar, no need to open or print the pdf...

.: Who is talking? .: What is the message? .: What is the support / who is the audience? .: Textbook - students (and teachers) .: Major conference publication - researchers students in the eld .: Work in progress, workshop - peers working on a specic topic
Did you already read many references, are they citing the main references within the eld.

Pass 2 - Get the authors message

Introduction + Conclusion .: What is the problem (look for However) .: How are the authors solving it .: What are the authors contributions Discussion .: What are the insights, the limitations, the relevance of the work.

mercredi 31 aot 2011-w

14

Pass 3 - careful reading

From beginning to end ... ... or backwards

.: Read actively, use pens .: Talk to other students .: Get help from related work papers if it gets too complicated .: Come back to it later

mercredi 31 aot 2011-w

15

If it gets too complicated move to another one that may be more general

Living with it

Research papers summarize months or years of work in a few pages Come back, re-read, question, re-discover.

mercredi 31 aot 2011-w

16

If the paper is core to what you are doing. Particularly relevant to your course project or masters thesis.

Plenty of other ways of reading...

Objectives .: Literature review .: Survey .: Reviewing articles for colleagues, journal or conferences Other sources: .: Books .: Patents .: Web content

mercredi 31 aot 2011-w

17