|
F2023
Presentation
|
|
|
Part 1: Presentation Topic
|
Due: October 13
|
Description
In
Part 2 of this presentation
You will be making a oral presentation of a paper in the field of Human-Computer Interaction,
This paper should
- be an academic paper and not a magazine article
- not be a review paper
- have been published at most 10 years ago so that it describes fairly recent advances in the field
- have been published at least 5 years ago so that its impact on its field can be assessed
In this presentation you will be explaining:
- The research results described in this paper and their significance
- How this paper impacted its field: what previous research was it built on, and how have this
paper's results been used since its publication?
Papers can be found at
https://library.torontomu.ca/.
A good starting place is the
ACM Digital Library
and its
SIGCHI - Special Interest Group on Computer Human Interaction
To find out the impact of a particular paper, you will need to see how it has been cited since
its publication. To do so, it is recommended that you use
Scopus.
If you need help with any of this research, it is recommended that you physically go to the
Ryerson library and seek assistance from the Computer Science
Subject Liaison Librarian.
Suggestions for how to choose a paper
The paper that you will present needs to be a paper that advanced its field.
Finding such a paper will require some investigation because this is a post-facto characterization:
authors don't necessarily know that they are advancing a field,
and even if they do, they don't necessarily know how.
They often have some idea about why what they do is important, and they may suggest how to continue the
research; however what really matters is what happens next when other people build on their work.
Sometimes other people will have some insights that the authors did not have themselves.
One indicator that a paper has some impact is whether it is cited a lot, but it is not the only indicator.
Other possible numeric indicators are whether some followup papers are cited a lot.
However, citation counts only measure quantity of activity, not quality.
Quality can be sometimes inferred by the peer-review process: when a paper is published in a more "reputable" venue,
there is a good chance that it is an important paper. However, this is still not a perfect way of determining
quality: all publications are run by people who often think in similar ways and therefore have preferences and biases.
Also, there are many factors that come into play when an author decides where to present a paper,
and these factors (such as turnaround time) often have nothing to do with the quality of the paper being presented.
Ultimately you will need to look at the content of the paper and what work was done in that field before and after
to determine for yourself whether it is advancing a field. This is the core of this presentation.
There is no single way to go about finding a good paper for this purpose,
but here is an approach that can often yield reasonable results:
- Choose a paper (or papers) that interest you.
It doesn't matter why they interest you - this is personal.
These papers do not have to be related to your area of research, but they can be if you want.
Do not worry about the age of the papers, because these will not necessarily be the papers that you present.
These papers will be the starting points of your search for a paper to present:
They will not necessarily be the papers that you end up presenting, but they can be.
They can also instead guide you towards the paper that you will present.
- Read these papers to try to figure out what is being proposed that is new in the field.
It could be a new artifact, but what is even better is a new way of doing things or a new way
of thinking about things. Let's call this a "new idea".
- Find out whether this new idea is really new, how much of it is new,
and whether the newness was already mostly developped prior to this paper,
or even whether somebody else came up with a better similar idea.
This is detective work.
You will need to look at the references to figure all of this out,
and sometimes you will need to look at which papers are citing the references,
or the papers themselves as well.
- Repeat until you find a paper that is actually advancing a field by proposing something new that
is subsequently worked on in a productive way that is appreciated not just by the authors of the paper
but by others in the field.
This process of reading papers may seem overwhelming at first as it involves a lot of reading.
The key to net getting bogged down is to learn to skim papers to decide whether they are relevant to your quest.
Again this is a not a process with 100% accuracy, but it helps. Here are good starting points to determine what a
paper is about:
- The title and keywords
- The authors if you know what other work they usually do
- The publication venue if you are knowledgeable about the type of work that they usually publish
- The Abstract should summarise the paper reasonably well if it is properly written
- Sometimes the Introduction, Conclusion, Reflections, and Future Work sections (or similar sections)
do a better job than the abstract.
- The Related Work and Reference section are very helpful for the detective work you need to do to choose a paper.
- You can often skip the Methodology and Usability Studies sections at first.
They are important to validate the ideas in the paper, but not to help you choose a paper.
Then what is left is the actual content describing the work done.
You will need to read it for papers that are a serious part of your detective work.
Hand In
Submit on D2L to the "CP8205 Presentation Topic" folder
- A 1-2 page report explaining what paper you will be presenting, why you have chosen it:
why its results are important and how this paper helped advance knowledge in its field.
- A copy of the main paper that you will be presenting
- Copies of other papers you intend to discuss in your presentation
This page is maintained by
Sophie Quigley
(cps613@cs.torontomu.ca)
Last modified
Tuesday, 22-Aug-2023 00:18:35 EDT