InformationCourse code and title
SCIE1000
Theory & Practice in Science
Semester
Semester 2, 2022
Type
Online, non-invigilated assignment, under ‘take home exam’ conditions.
Technology
File upload to Blackboard Assignment
Date and time
Your assignment will begin at the time specified by your course coordinator. You have a
fixed 12-hour window from this time in which it must be completed. You can access and
submit your paper at any time within the 12 hours. Even though you have the entire 12
hours to complete and submit this assessment, the expectation is that it will take
students around 2 hours to complete.
Note that you must leave sufficient time to submit and upload your answers.
Permitted materials
This assignment is closed book – only specified materials are permitted.
Ensure the following materials are available during the available time:
Recommended
materials
Instructions
•
The SCIE1000 lecture book, workshop activities & solutions, and your personal
notes from the course are permitted.
•
UQ approved calculator; bilingual dictionary; phone/camera/scanner
You will need to download the question paper included within the Blackboard Test. Once
you have completed the assignment, upload a single pdf file with your answers to the
Blackboard assignment submission link. You may submit multiple times, but only the last
uploaded file will be graded.
You can print the question paper and write on that paper or write your answers on blank
paper (clearly label your solutions so that it is clear which problem it is a solution to) or
annotate an electronic file on a suitable device.
Given the nature of this assessment, responding to student queries and/or relaying
corrections during the allowed time may not be feasible.
Who to contact
If you have any concerns or queries about a particular question or need to make any
assumptions to answer the question, state these at the start of your solution to that
question. You may also include queries you may have made with respect to a particular
question, should you have been able to ‘raise your hand’ in an examination-type setting.
If you experience any interruptions during the allowed time, please collect evidence of
the interruption (e.g. photographs, screenshots or emails).
If you experience any technical difficulties during the exam, contact the course
coordinator a.penton@uq.edu.au. Note that this is for technical difficulties only.
Late or incomplete
submissions
In the event of a late submission, you will be required to submit evidence that you
completed the assessment in the time allowed. This will also apply if there is an error in
your submission (e.g. corrupt file, missing pages, poor quality scan). We strongly
recommend you use a phone camera to take time-stamped photos (or a video) of every
page of your paper during the time allowed (even if you submit on time).
Page 1 of 18
If you submit your paper after the due time, then you should send details to SMP Exams
(exams.smp@uq.edu.au) as soon as possible after the end of the time allowed. Include
an explanation of why you submitted late (with any evidence of technical issues) AND
time-stamped images of every page of your paper (e.g. screen shot from your phone
showing both the image and the time at which it was taken).
Academic integrity is a core value of the UQ community and as such the highest
standards of academic integrity apply to assessment, whether undertaken in-person or
online.
This means:
Further important
information
•
You are permitted to refer to the allowed resources for this assignment, and you
must not use any instances of work that has been submitted previously
elsewhere.
•
You are not permitted to consult any other person – whether directly, online, or
through any other means – about any aspect of this assignment during the
period that it is available.
•
If it is found that you have given or sought outside assistance with this
assignment, then that will be deemed to be cheating.
If you submit your answers after the end of allowed time, the following penalties will be
applied to the total mark available for the assessment:
•
Less than 5 minutes – 5% penalty
•
From 5 minutes to less than 15 minutes – 20% penalty
•
More than 15 minutes – 100% penalty
These penalties will be applied unless there is sufficient evidence of problems with
the system and/or process that were beyond your control.
Undertaking this online assignment deems your commitment to UQ’s academic integrity
pledge as summarised in the following declaration:
“I certify that I have completed this assignment in an honest, fair and trustworthy
manner, that my submitted answers are entirely my own work, and that I have neither
given nor received any unauthorised assistance on this assignment”.
Page 2 of 18
Semester Two Assessment, 2022
SCIE1000
To answer each question you will need to use the information on pages 17-18. Your
solutions will be marked on the correctness and clarity of your explanation and communication. Include units in your answers wherever relevant.
Each question is graded on a 1–7 scale with the last part of the question being at an
advanced level which must be attempted for students aiming for a grade of 6 or 7.
Question 1.
When designing certain types of scientific studies, researchers must predict the infrastructure needs
to achieve certain goals. In [1] some important considerations were the number of movement sensitive
cameras required to cover the area that they wished to survey and also how to best process the image
data. Consider a situation wherein each of the 225 cameras in the survey area will trigger 5840 times
over the course of the 4 year study.
(a) If a computer would take on average 20 seconds to classify each image, how long, in days, would
it take a single computer to classify the entire collection of images?
(2 marks)
(b) As an alternative to a single computer classifying all the images, the power of citizen science
can be utilised instead, having 68000 volunteers classify the images. Suppose that each image is
classified by 9 different people to ensure accuracy, and it takes each person 2 minutes to classify
each image on average. If each person classifies the same number of images, how long, in days,
will it take each person to classify their images?
(2 marks)
(question continued over)
Page 3 of 18
Semester Two Assessment, 2022
SCIE1000
(c) When considering the time taken to classify the images in the study described in [1], is utilising
citizen science beneficial? Briefly justify your answer.
(1 mark)
(d) (Advanced) The UQ Faculty of Science has started a new initiative encouraging the public to
engage in citizen science projects. You have been asked to write a short paragraph as a part of
an article to be published on the faculty website, discussing the benefits of citizen science and
why people should be getting involved. Using details provided on the information sheet, write a
short paragraph (only three or four sentences) addressing this issue. Your communication style
should be appropriate for the level of knowledge of a typical member of the public. (2 marks)
(next question over)
Page 4 of 18
Semester Two Assessment, 2022
SCIE1000
ln(Number of distinct species)
Question 2.
Swanson highlights the strengths of large scale imagery in animal observations in [1]. The area under
observation was separated into cells, where each cell is the area surveyed by one camera (∼ 5 km2 ).
The figure below is adapted from observations taken by Swanson on 20/09/12. It plots the total
number of distinct species found as a function of the number of cells surveyed, as a log-log plot.
Species data
3.0
2.5
2.0
1.5
0.0
0.5
1.0
1.5
2.0
ln(Number of cells)
2.5
3.0
(a) Develop a linear equation to fit the data shown in this graph. To simplify writing your equation, let y represent the (unitless) natural logarithm of the number of distinct species, and let x
represent the (unitless) natural logarithm of the number of cells. Make sure to explicitly write
down your final result for the linear equation and draw your line of best fit on the figure. (3 marks)
(question continued over)
Page 5 of 18
Semester Two Assessment, 2022
SCIE1000
(b) Develop an appropriate model for species count (S) as a function of the number of cells (A),
making sure to explicitly evaluate all the numerical constants that you can.
(2 marks)
(question continued over)
Page 6 of 18
Semester Two Assessment, 2022
SCIE1000
Number of distinct species
(c) (Advanced) Given that the species area curve is collated using data from across the surveyed
area, one might expect to get similar results from day to day. However, this was not the case
for the observations in [1]. The figure below plots the raw data for the figure from part (a) with
some additional data from a day approximately 3 months later. Describe how the model for the
data represented by red crosses would differ from your answer in part (b) and comment on what
physical reasons there might be for this difference.
(2 marks)
20
15
10
5
Species data 20/09/12
Species data 28/12/12
0
0
5
10
15
Number of cells
20
25
(next question over)
Page 7 of 18
Semester Two Assessment, 2022
SCIE1000
Question 3.
(a) The information sheet provides some data about the number of zebras observed by a particular
camera. Identify which of the mathematical models discussed in class (power, periodic, exponential, surge, etc.) would best describe the number of zebras observed by this particular camera as a
function of time. Use the data on the information sheet to develop a suitable mathematical model,
including numerical values for the appropriate constants, for the number of zebras observed by
this particular camera each day as a function of time.
(3 marks)
(question continued over)
Page 8 of 18
Semester Two Assessment, 2022
SCIE1000
(b) You have been asked to participate in some outreach at a local high school and you are posed this
question when showing the model from part (a):
I’ve seen videos on TV where the zebras are just everywhere in Africa, why
are there sometimes none and sometimes lots?
— Year 8 student
Write a short paragraph (2-3 sentences) suggesting the physical reason for the shape of your model
from part (a). Your communication style should be appropriate for the level of knowledge of a
typical high school student.
(2 marks)
(c) (Advanced) Using the image classification data from [1] we can plot the number of zebras that are
seen by a single camera over an approximately 10 month period.
There are likely some differences between the model that you created in part (a) and the data
shown here. Identify any differences between the collected data and your model from part (a) and
suggest a physical reason that may cause these differences to occur.
(2 marks)
(next question over)
Page 9 of 18
Semester Two Assessment, 2022
SCIE1000
Question 4.
Understanding accuracy in classification is an important consideration in large scale studies, especially
those that are using techniques such as machine learning or citizen science. This is usually done by
giving each classifier (in the case of [1] this is the everyday people that are helping to classify the
images) a set of data to classify, and then giving this same data to one or more experts to do the same.
This creates a “gold standard” to compare against.
(a) Using the data supplied on the information sheet, is it possible to evaluate the effectiveness of the
citizen scientists’ image classification? Explain your answer.
(1 mark)
(b) In addition to the information provided, suppose that you are also told that out of all images that
were classified as containing an animal, the probability that the image didn’t contain an animal
was 3%. Using a binary classification table or tree diagram, compute the sensitivity and specificity
of the citizen scientists’ classifications.
(3 marks)
(question continued over)
Page 10 of 18
Semester Two Assessment, 2022
SCIE1000
(c) Based on your answer to (b), are the citizen scientists a good way to conduct image classification
on the Serengeti? Explain your answer.
(1 mark)
(d) (Advanced) Someone wrote to the scientist running the Snapshot Serengeti project and asked the
following:
In the age when artificial intelligence exists, why are you utilising citizen scientists? There are publicly available classifiers that could sort
through the images and would surely be more accurate than just ordinary people.
— Sir Jonathan Science II
The Snapshot Serengeti team did a study on the performance of computer vision on this data
set. Out of the images that contain no animals, the algorithm classified an image as containing
an animal 3.5% of the time. They also found that out of all the times the algorithm returned
the result of an animal being present, it was wrong 11% of the time. Calculate the accuracy of
the method suggested by Sir Jonathan Science II and write a short paragraph (3-4 sentences) in
response to them, explaining the result and possible reasons for it.
(2 marks)
(next question over)
Page 11 of 18
Semester Two Assessment, 2022
SCIE1000
Question 5.
Understanding the interaction between populations of different species is an important focus of surveys
such as that undertaken in [1]. Due to the vast number of different interactions, it is often very difficult
to model ecosystems in their entirety. However, it is possible to develop simplified models to understand
some interactions between species. Let G(t) be the number of gazelles and L(t) be the number of lions,
at time t in years. The following equations model the interactions between the gazelles and lions.
G′ = 0.15G − 0.07LG
L′ = −0.3L + 0.0025LG
(a) What are the physical significance and units of the constants 0.15 and −0.3 in these equations?
(1 mark)
(b) What does the model predict would happen to the gazelle population if this interaction stopped,
for example if lions went extinct?
(1 mark)
(question continued over)
Page 12 of 18
Semester Two Assessment, 2022
SCIE1000
(c) In the area surveyed, there are approximately 150 000 gazelles and 1000 lions. Using 150 and 1 (in
units of thousands of animals) as your respective starting values, utilise 2 steps of Euler’s Method
to calculate the predicted populations of these species in 3 years (step size of 1.5 years).
(3 marks)
(d) (Advanced) Using what you know about models from the philosophy of science component of
SCIE1000, explain how the nature of the above model might limit its use in predicting species
populations and how, despite these limitations, the model could nevertheless be of use. (2 marks)
(next question over)
Page 13 of 18
Semester Two Assessment, 2022
SCIE1000
Question 6.
The following is a sample of a code that was supplied to citizen scientists participating in the image
classification effort.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
from pylab import *
# Initialise arrays
All_IDs = zeros(10)
All_Animals = zeros(10)
mean_encounter_rate = 0.246
no_encounter = 0
# Begin loop
for i in range(10):
Photo_ID = int(input(“Please enter the photo ID: “))
# The codes for animals are: 1-zebra, 2-lion, 3-antelope, 4-wildebeest
Observed_Animal_ID = int(input(“Please enter the ID for the animal that you’ve
spotted (0 if no animals are spotted): “))
16
17
if Observed_Animal_ID == 0:
18
no_encounter = no_encounter + 1
19
20
All_IDs[i] = Photo_ID
21
All_Animals[i] = Observed_Animal_ID
22
23 if no_encounter