Department of Computing and Information Systems
The University of Melbourne
COMP90049 Knowledge Technologies, Semester 2 2019
Project 1: Word Blending in Twitter
Released: Friday 16 Aug
Due: Research Paper: Friday 13 Sep – 5PM
Reviews: Wednesday 18 Sep – 5PM
Marks: The project will be marked out of 20 (according to the given
criteria), and will contribute 20% of your total mark.
The goal of this project is to develop and critically assess methods for detecting word blends
among frequent terms in Twitter data, and to express the knowledge that you have gained about
this task in a short research paper. Twitter users use language innovatively, and coining new
terms by blending two existing words is a common phenomenon, known as lexical blending.
Consider the following examples:
Component 1 Component 2 Blend word
Britain exit Brexit
spoon fork spork
breakfast lunch brunch
You will detect occurrences of blend words among a pre-processed list of tokens from a
Twitter data set, using a reference set of English words from a dictionary, and using methods
for approximate string matching as encountered in the lectures. We will also provide you with a
set of tweets the token list was extracted from, which you may (but are not expected to) use.
You will evaluate the output of your algorithm(s) against a list of true word blends. The project
aims to reinforce concepts in approximate matching and evaluation, and to strengthen your
skills in data analysis and problem solving.
The goal of this assignment is not to develop a system which achieves near-perfect precision
(in fact, this is impossible – we are developing knowledge technologies after all!).
1. One or more programs, implemented in the programming language(s) of your choice, which
Process the data input file(s), to identify word blend candidates
Identify word blend candidates among a set of tokens, with the help of a reference
collection of tokens (dictionary)
Evaluate the matches, with respect to the list of true word blends, using one or more
2. A README that briefly details how your program(s) work(s). You may use any external
resources for your program(s) that you wish: you must indicate these, and where you obtained
them, in your README. The program(s) and README are required submission elements,
but will not typically be directly assessed.
3. An anonymous short research paper of 1100–1350 words (±10%), as a single file in PDF
format, which should include:
A short description of the problem and data set
A brief summary of some relevant literature
A brief explanation of the approximate matching techniques used
Presentation of your results in terms of the evaluation metrics discussed and illustrative
A discussion on the knowledge you have gained about the problem at hand, and about the
(un)suitability of the approaches you have adopted
4. Reviews of two research papers written by your peers, each of 250-350 words (±10%),
comprising 4 out of the 20 marks and a critical self-reflection on your own work.
The Twitter dataset is based on the data set presented in
Jacob Eisenstein, Brendan O’Connor, Noah A. Smith,and Eric P. Xing. 2010. A
latent variable model forgeographic lexical variation. In Proceedings of the 2010
Conference on Empirical Methods in Natural Language Processing (EMNLP 2010),
You need to cite this paper in your research paper.
The list of blend words was compiled using resources presented in the following
Deri, A. and Knight, K. (2015) How to Make a Frenemy: Multitape FSTs for
Portmanteau Generation. In Human Language Technologies: The 2015 Annual
Conference of the North American Chapter of the ACL, pages 206–210
Das, K. and Ghosh, S. (2017) Neuramanteau: A Neural Network Ensemble Model for
Lexical Blends. In Proceedings of the The 8th International Joint Conference on
Natural Language Processing, pages 576–583
Cook, P. and Stevenson, S. (2010) Automatically Identifying the Source Words of
Lexical Blends in English. In Computational Linguistics, Volume 36(1)
You need to cite these papers in your research paper.
You are strictly forbidden from reproducing documents in the document collection in any
publication, other than in the form of isolated examples.
Additionally note that the document collection is a sub-sample of actual data posted to Twitter,
without any filtering whatsoever. As such, the opinions expressed within the documents in no
way express the official views of The University of Melbourne or any of its employees, and my
using them does not constitute endorsement of the views expressed within. We recognize that
some of you may find certain of the documents in bad taste and possibly insulting, but please
look beyond this to the task at hand. The University of Melbourne accepts no responsibility for
offence caused any content contained in the documents.
(1) Short research paper: (15 marks out of 20)
Method: (30% of the paper mark)
You will make one or more suitable hypotheses regarding the coinage of blend words, and
design experiments using one or more approximate matching methods which could
plausibly test your hypotheses. You will use the data to evaluate the method(s) logically
and formally. You will describe your implementation in a manner that would make your
Critical Analysis: (40% of the paper mark)
You will analyze the effectiveness of your system(s), referring to the underlying
theoretical behavior where appropriate. You will attempt to confirm or reject your
hypotheses, using supporting evidence in terms of illustrative examples and evaluation
metrics. You will derive some knowledge about the problem of identifying the causes of
Report Quality: (30% of the paper mark)
You will produce a report which is commensurate in style and structure with a (short)
research paper. You will express your ideas clearly and concisely, and remain within the
word limits. You will include a short summary of related research.
NOTE: A marking rubric is available on LMS to indicate what we will be looking for in each
of these categories when marking.
(2) Reviews and self-reflection (5 marks out of 20)
You will have 250–350 words to respond to three “questions” for two research papers of your
peers (2 marks each) and for your own paper (1 mark):
• Briefly summarize what the author has done
• Indicate what you think the author has done well, and why
• Indicate what you think could have been improved, and why
Completing the reviews is expected to take about 3–4 hours in total.
Changes/Updates to the Project Specifications
If we require any (hopefully small-scale) changes or clarifications to the project specifications,
they will be posted on the LMS. Any addendums will supersede information included in this
For most people, collaboration will form a natural part of the undertaking of this project.
However, it is still an individual task, and so reuse of ideas or excessive influence in algorithm
choice and development will be considered cheating. We will be checking submissions for
originality and will invoke the University’s Academic Misconduct policy
(http://academichonesty.unimelb. edu.au/policy.html) where inappropriate levels of
collusion or plagiarism are deemed to have taken place.
Late Submission Policy
You are strongly encouraged to submit by the time and date specified above, however, if
circumstances do not permit this, then the marks will be adjusted as follows:
Each business day (or part thereof) that this project is submitted after the due date (and time)
specified above, 10% will be deducted from the marks available, up until 5 business days (1 week)
has passed, after which regular submissions will no longer be accepted.