Hi, I am a biomedical imaging and visualization researcher who investigates how the study of brain connectivity and machine perception can help advance the understanding of biologically inspired artificial intelligence.
Counts as 4 quizzes!
Now in Blackboard!!
# How to read a paper ## So many papers.. * you can not never read them all ## When to read papers? * special topic ** you grab a single individual paper based on the title and content * literature/research state-of-the-art ** every paper has some related work ** but you can not possible read it all ** you need to skim it * state of the art report (STAR) / review paper ** include hundreds of papers and you can not read them all ** STAR are great summaries for a specific research subfield or topic * paper review ** reviewing is giving feedback about some submitted work ** you decide if the paper is scientifically sound and should be accepted ** 3 reviewers who then together with an editor or PC member decide on the acceptance ** great experience for your resume/CV ## What is a good paper? * abstract, figures, conclusion ** abstract is the summary / the overall story ** ** example 1: dojo ** ABSTRACT *** first 4 sentences in abstract give background *** then contribution is mentioned (2 new softwares Mojo and Dojo) *** evaluation is mentioned as a between-subjects quantitative user study (objective based on scientific hard facts) **** not in the abstract: 30 people total without any experience, 10 per software to compare the user performance **** qualitative feedback (subjective) by asking users how they liked it ** TEASER *** we know a little bit more now and there is some repetition ** FIGURES *** Fig 2 gives more background and highlights the importance *** because we want to convince the reviewer about our contribution *** Fig 3 shows one of the tools: Raveler *** Fig 4 shows examples of errors and some feature of Dojo *** Fig 5 shows 3rd tool Mojo *** Fig 6 shows Dojo: web-based and multi-user *** Fig 7 shows example of merge workflow *** Fig 8 shows example of split workflow *** Fig 9 shows example of adjust workflow *** Fig 10 shows collaborative view / multi-user *** Fig 11 shows some example of proofreading in 3D ** CONCLUSION *** some more repetition about the content of the work *** some limitations we did not know about yet *** future ideas HOW TO READ A PAPER ** 1. read abstract to get outline of the story *** but you will still have a lot of questions ** 2. look at the images *** start with the teaser if it exists *** go figure by figure ** 3. go to the end of the paper and read conclusions ** 4. look at scientific evidence (maybe look at plots first) ** example 2: scalable interactive visualization
# How to write a paper ## Author order ** follows U-shape with first and last author being the most important ** (only in CS theory its different and alphabetical order) ## Overlead LaTeX/Templates * MICCAI ** each conference and each journal have their templates ** use overleaf templates if exists ## Outline (first step) * Introduction * Related Work * Method * Evaluation * Discussion and Results * Conclusions and Future Work ## Add Figures.. ## Bullets ## Paragraphs
not sure if this is a good idea...
CS480 Lecture 40
By Daniel Haehn