Heidi`s Slides on 3.44 - Computer Science and Software Engineering

Download Report

Transcript Heidi`s Slides on 3.44 - Computer Science and Software Engineering

3.44
Heidi Newton
What I am here to talk about
• I have had a look at some of the 3.44
submissions.
• I am going to share some of my observations
from looking at them.
The annoying fine text nobody reads
• I was NOT a marker
• This talk is about the student work that was
submitted. The marking processes are outside
the scope of the talk.
Agenda (To ensure we cover
everything…)
•
•
•
•
•
•
•
Common issues overall
Artificial Intelligence (Intelligent Systems)
Complexity and Tractability
Formal Languages
Network Communication Protocols
Software Engineering
Computer Graphics and Vision
• And hopefully we can have discussions on each
topic.
Common Issues Observed
• Only covering 1 topic
• Only covering 1 topic well and doing poorly on
the other
• Covering stuff that is not Computer Science
• Covering other Computer Science topics
• Writing 14 pages of achieved level material
• Copying/ paraphrasing wikipedia
• Lack of personalisation – need student voice
• Jargon used incorrectly (or jargon defined using
jargon)
What does “Problem” mean in
Computer Science?
• There was a few misunderstandings here…
• It does not mean a bad thing
– Infact, computer scientists LOVE problems!
• Tasks (often interesting) that have to be
solved, not things that get in the way.
• Achieved: “describing key problems that are
addressed in selected areas of computer
Science”
There isn’t always a right and wrong
answer
• From my observations, and discussions over the
last few days, some students find this difficult.
– Excellence often doesn’t have right and wrong
answers
– Try to encourage them to have an opinion
– And to not be afraid to share it in their report with
explanations to explain and/or justify it
– It is okay if the person next to you has a different
opinon
– And it is even okay if the marker has a different
opinion, as long as you explained yours!
A few things about the report…
• 3.44 is a report, not an essay, not a narrative
• Some bullet points are fine (although use
sentences; not two or three words)
• A heading in the ugly format used above
would NOT lose marks as the marking is on
the content (but I would still avoid word art
)
• You are even allowed to submit a short video
instead of a report
Artificial Intelligence
• By FAR the most popular topic!
• Some VERY long reports (7 pages…) that ticked
the boxes for achieved MANY times but did
not touch Merit, so got Achieved
• Difficulties knowing what to say to chatbots
for Turing Test
• Not linking back to practical applications
(essential for excellence).
• Some applications discussed really weren’t AI
Suggestions for Improving Artificial
Intelligence
• More guidance for students on good questions and things to say in
the Turing Test
– Perhaps students would benefit from teachers doing a demo in class?
– Class discussions about how to talk to the chatbots are recommended
• We’re going to make some improvements to the field guide… (and
hopefully add 2 more topics for artificial intelligence).
– Reduce length of chatbots/ turing test activity.
– Going to post a discussion on the NZACDITT mailing list for thoughts
on this.
• Final note: remind students to keep conversations they put in their
reports appropriate
– If Cleverbot is saying offensive things (it does sometimes…), censor it.
– (Offensive material in work submitted to be externally marked is not
allowed according to NZQA rules)
Complexity and Tractability
• Field guide followers actually seemed to do
quite well in this section.
• A lot of wikipedia paraphrasing that did not
convey understanding
• Explanations with mathematical notation and
“n” without evidence the student knew what
they were talking about
• Some problems chosen lacked real world
applications
– Towers of Hanoi is an interesting puzzle (not really
a “problem”), but lacks real world applications
Suggestions for Complexity and
Tractability
• Follow the field guide or come up with your own activity,
rather than using wikipedia.
– Personalised examples are REALLY good
– Try and ensure there are real world applications for the problem
• If mathematical explanations about “n” are included in the
report, putting some numbers to it to explain it is more
convincing that the student actually understands the math
• Suggestions for problems:
– Travelling Salesman Problem
• Craypot variant
• Map of the school and the problem of delivering messages to
classrooms),
– Bin Packing (or the knapsack problem)
– Timetabling
– Factoring product of two primes (applications in encryption)
Formal Languages
• Personalization was a bit limited; students did
not really come up with their own state
machines
• Few examples of regular expressions
– An excellent idea from Chch is getting students to
make a database and use regular expression
queries
• Some students muddled this with
programming languages (which is in level 1)
• Not many students did this topic, I have not
seen very many examples of it
Network Protocols
• This is NOT an infrastructure standard!
• Rather, what are the jobs higher layer network
protocols need to do?
• Ensure reliability, security, acknowledge when
data is received…
• We are starting to think about how to deliver
this within the field guide…
• (e.g. Tablets of Stone)
Software Engineering
• Second most popular topic overall
• Industry talks were very popular
– Learn methodologies from field guide/ other resources
– Link to practical applications from what speaker says
• I’ve not looked at much of the work done in this area
• Tough to personalize.
– Essay style ok for this topic
– Use the student voice!!!
• I have heard a few interesting comments/ concerns
from teachers about it
– Anybody have anything to share?
• Explaining and discussing waterfall vs agile is a good
idea
• Don’t do it on their own program!
Graphics and Computer Vision
• The most mathematical topic
– Understanding matrices is a big part of it
• This is not about using graphics programs or
using graphics
• Examples and widgets within the field guide
provide good examples to focus on.
• Focusing on the error control coding of QR codes
does not meet the standard (It’s level 2)
• There were not many students that did these
topics, and I have not looked at them (not my
Important Messages
• If you are not confident and/or do not have
your own resources, use the field guide or
other good resources.
• Stay well away from wikipedia and field guide
paraphrasing. It does not demonstrate
understanding.
• 14 pages of text and no images might not
show a good understanding
• Don’t give them templates, student voice!
– Starting points, questions, checklists, OK, but not