2. LEIC 2000, MEIC 2006, DEIC 2014
Visiting Student at MIT in 2009
Lecturer at Técnico since 2002
Visiting Scholar at TUM in 2017
First publication in 2004
Technical/research interests:
Cybersecurity applied to
Internet of Things and
Cloud Computing
http://web.tecnico.ulisboa.pt/miguel.pardal
5. … wait for the next meeting to ask questions
… miss meetings
... arrive late
… ignore what your supervisor says
Adapted from slide by Hugo Nicolau
6. … ask questions
… argue back
... keep up the pace – no weeks off
... read as much as you can
... take notes and draw sketches
... have new ideas
Adapted from slide by Hugo Nicolau
7.
8. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How write technical papers
9. What is the problem?
Where is the value?
Who are the stakeholders?
How can we solve the problem?
In a new or better way
In a repeatable way
10. Most web pages are NOT scientific papers
Including Wikipedia
Why?
What makes a document a scientific paper?
Its results were reached following the
scientific method:
Hypothesis
Experiments
Conclusions
Peer-reviewed
A peer is another scientist or expert
11. “X is better than Y!”
How do you know that?
Based on your own results, or
based on the work of others
Accepted by the scientific community
Citations of previous work
12. Examples:
Cite to support a statement
“System XPTO was evaluated as the fastest [7]”
Cite to introduce a new system/result
“Smith et al. [22] proposed a new system to…”
[7] and [22] above are references to the 7th
and 22nd papers cited in the bibliography
How to name the authors in text
1: Smith
2: Smith and Williams
3+: Smith et al.
13. Our research is only possible because of
the work of others before us
18. Citations
How many papers from other authors
cite papers by the author
Hirsch-Index (h-index)
Attempts to measure both the
productivity and citation impact
For instance, an h-index of 17 means
that the scientist has published at
least 17 papers that have each been
cited at least 17 times
i10-index
Number of publications with at least 10
citations
19. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How to write technical papers
29. Start with most cited papers
They are the ones the community is reading and
using in their work
Also see the most recent papers
from top sources
Published in the last 3 years
These recent papers will cite previous relevant work
In ACM, IEEE, Usenix, … or others (ask your advisor)
30. Good conference ?
Accept rate < 20%
Only 1 in 5 submitted papers gets accepted
Check conference rankings
CORE, GGS
Good journal ?
Check journal rankings (Quartile 1 are the best)
Scimago
Impact factor
Average number of citations to
recent articles published in the journal
36. Write a short summary of paper
Should fit in an index card
What is relevant?
System characteristics
Assumptions made by authors
Examples/scenarios/use cases
used for evaluation
...
State strong and weak points
37. You can use note taking apps, with search
OneNote, EverNote, …
https://www.onenote.com/
https://evernote.com
39. From the paper you selected
Or from a random paper in the magazine
In 5 minutes
2 minutes to read
3 minutes to write
Write down some notes about it
Contributions
Strengths, Weaknesses
Points of interest
Comparison
40. Tables are great for comparing things
Start with columns from existing table, or
think of your own
Add more columns if you need
Empty cells show what you don’t know yet
…
Latest Version
Initial Release
System
10.0
1983
Windows
[Gates83]
10.11 (El Capitan)
1984
Mac [Jobs84]
1991
Linux [Torvalds91]
43. Review and improve your notes
Contribution
Strengths & Weaknesses
Points of interest
Examples, use cases, diagrams, plots, …
Compare with your work
Study details in depth
As needed for your work
Come back later for more
As your work matures, reading a related work
paper again can provide more insight
44. 1000’s 100’s 10’s
• Keywords
search
• Quick read
• Discard/save
• Normal read
• Take initial notes
• Save PDF and
BibTeX
• Add to
comparison table
• Full/deep read
• Improve notes
• Complete
comparison table
(including a row for
your work)
• Cite in
paper/dissertation
45. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How to write technical papers
46.
47. Move 1 – Establish the “territory”
Move 2 – Establish the niche
Move 3 – Occupy the niche
Your advisor is very important for these moves
48. Move 1 – Establish the “territory”
Claiming centrality and/or
Making topic generalizations and/or
Reviewing items of previous research
Move 2 – Establish the niche
Counter-claiming or
Indicating a gap or
Question-raising or
Continuing a tradition
Move 3 – Occupy the niche
Outlining purposes or
Announcing present research
Stating principle findings
Indicating research article structure
http://www.wikihow.com/Write-a-Swalesian-Introduction
49. Context
Who are the entities involved?
Goals / motivations
Who are the stakeholders?
E.g. customer
50. A solution design can be represented in diferent
perpectives
Different abstraction
Different detail levels
Conceptual
Logical
Physical
51. Concepts
Information entities (ex. Order, Complaint)
Processes (ex. Accept Order, Respond to Complaint)
From the problem, extract solution
requirements
Functional
Quality (Performance, Security, Reliability, …)
Need to be accepted by the stakeholders
53. Formal description of system
System specification
E.g. objects
Structure
Interfaces
Components
Behavior
Events
Expected responses
54. A model can be textual
Example: attacker model
or visual
Example: UML package diagram
Represent structure and behavior
Each diagram is a perspective on the system
55. A model is a simplification/abstraction of
a complex object
To show relevant characteristics
Easier to manipulate than the actual object
58. … it achieves goals and perspectives are aligned
Implementation passes tests that show that the system
follows the specification that, in turn, satisfies the
stakeholder’s requirements
Conceptual Logical
Physical
requirements specification
implementation
translate
detail
59. Regarding your thesis work
Or the system chosen earlier
In 5 minutes
Draw a structural diagram, representing the
main components of the solution
60. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How to write technical papers
61.
62. Ask questions
Design experiments to find answers
Possible, ethical, rigorous, reproducible
Perform experiments and collect data
Interpret data
Are the answers satisfactory?
If not, what are the follow-up questions
Redesign if needed and do more experiments…
63. How long, on average, does the system
take to respond to requests?
How many simultaneous users can the
system hold without degrading
performance levels?
…
64. Qualitative
Answer questions about the
quality of something
Are requirements satisfied? Fully? Partially?
Quantitative
Answer questions about the
quantity of something
Use metrics, key performance indicators
65. Regarding the paper you selected
Or another one in the magazine paper
In 5 minutes
Pick one figure with a plot
What are the questions being answered by that
figure?
66. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How to write technical papers
67.
68. The work is almost complete
Is writing worth it?
YES
To write is to express thoughts into words
So that others may learn
Your own learning is not complete without the
“distillation process” that comes with writing
69.
70. Use the most common science language,
so that your work may reach the widest
audience
English
What about Portuguese and other languages?
Still important for scientific divulgation
72. 3 tells
Say what you will say
Say
Say what you said
Provide guidance and context to the reader
73. Each “middle” chapter in a dissertation
should have:
Introduction text
Main content sections
Summary section
Highlight most important points to carry forward
Transition to next chapter
74. Most important ideas first
Should appear in the beginning of the
paragraph
Details and alternatives should appear later
75. Be direct, to the point
“There are many important systems in computer
history, in particular, regarding remote graphical
systems, one of the first widely used contributions was
the X system [22]”
->
“The X system [22] was the first widely used remote
graphical system. It was an important system in
computer history.”
76. Use more formal language
Avoid oral contractions
We’re -> we are
Don’t -> do not
Gonna -> going to
Ain’t -> I am not
Avoid possessives
Joana’s work
-> The work by Joana
The system’s characteristics
-> The characteristics of the system
77. Use correct tense
Proposal
Future tense – will do task
Dissertation
Past tense – did task
Present tense – something is available now
Use the third person
Sometimes first person, plural is OK – we
78. Avoid passive voice
It creates uncertainty about who is the
subject of the phrase
“The system was shown to have good performance”
vs
“We have shown that the system has good
performance”
79. Avoid informal (colloquial) expressions
These are bad sentences:
“At the end of the day, the best system is …”
“It is a matter of life and death that…”
“There’s more than one way to skin a cat, so a
different approach was attempted…”
“The performance has hit a wall…”
82. Avoid possibly rigorous terms used in a
non-rigorous way
“scalable”
It means that the system can sustain a
performance level when the number of users
increases by orders of magnitude (10, 100, 1000)
Not that it supports many users
“real-time“
It means subject to specific time constraints
Not that the system is fast
83. Avoid quotes
Example: The “advanced” option is …
Conveys imprecision, lack of rigor
Avoid et cetera
Example: The system modules include the
graphical interface, the business logic, etc.
Conveys lack of rigor, again
If you are going to list, list everything
Use categories instead
84. A dissertation is also a story
It has a beginning – the problem
It has an end – the solution
It should answer the following questions:
What is the problem being solved?
What are the requirements?
What is the background and related work?
What will be done and how will it contribute to a
solutions to the problem?
What are the technical challenges?
How will the results be evaluated?
88. Process to increase the quality of the
scientific work and writing
Assure sound science practices were followed
Guarantee writing is clear
Common variations:
Blind review
You will not know who the reviewer is
Double blind review
The reviewer also does not know who you are
89. English quality
Structure
Literature review
Complete and to the point
Motivation for decisions
Rigor
Avoid misunderstandings
Identify the limitations of your own work
They will be found anyway (sooner or later)
They are future work opportunities
90. Reviews are not always constructive…
Do not get upset
The comments are about your not, not about you
You should be your worst critic
The first to know what are the things to improve…
Keep calm and carry on
91. What is research
How to survey related work
How to propose solutions
How to evaluate solutions
How to write technical papers
92. Science is personal and social
People are central
Survey the related work
Learn from other authors
Cite them
Compare what they did
Writing about your work
Explain in a concise way
Learn from the reviewers
Share with others
93.
94. IEEE Authorship Series and toolbox:
http://www.ieee.org/publications_standards/publications/
authors/authors_journals.html
Guide by Miguel and Joana Pardal:
http://web.tecnico.ulisboa.pt/miguel.pardal/www/doc/
quick-guide-research.pdf
Thanks
Miguel P. Correia for the review
Hugo Nicolau for sharing his ideas
95.
96. Widely used
Produces documents with excellent aesthetics
Easy to follow template rules
Easy to reuse text in different templates
Built-in support for mathematical expressions.
Generates the reference list automatically!
BibTeX
It is a programming language (with comments)
These are the main contributions:
% do not forget to mention the dataset!