SlideShare a Scribd company logo
1 of 32
OpenEssayist: Extractive Summarisation
and Formative Assessment
of Free-Text Essays
Nicolas Van Labekeļ‚, Denise Whitelock ļ‚,
Debora Field ļ‚‚, Stephen Pulmanļ‚‚, John Richardson ļ‚
ļ‚ Institute of Educational Technology ā€“ The Open University
ļ‚‚ Department of Computer Science ā€“ University of Oxford
SAFeSEA: Research Questions
ā€¢ How can an automated system detect passages
on which a human marker would usually give
some feedback ?
ā€¢ Can existing methods of information extraction,
summarization be adapted to select content for
such feedback ?
ā€¢ How effectively can these methods deliver
feedback ?
ā€¢ What effect does these techniques have on essay
improvement? On current essay and in future
ones ? On self-regulation and metacognition ?
Context
ā€¢ Essays: Open University (UK) postgraduate assignments
ā€“ Distance learning, adult learners
ā€“ 1500+ words, free-text & open-ended questions
ā€¢ No ā€œGold Standardā€, wide range of content
ā€“ Perfect test ground for extractive techniques
ā€“ Impact of lack of (or limited) domain knowledge?
ā€¢ Bulk of activity (i.e. writing) takes place outside system
ā€“ Usage of drafts ā€œvaries a lotā€ among students
ā€“ Nature, scope and timing of feedback?
ā€¢ Limited possibility for ā€œmockā€ experiments:
ā€¢ testing & evaluation on ā€œliveā€ material
ā€¢ Connection with summative (tutor-based) assessment ?
Education Postgraduate Course H810
Accessible online learning: supporting disabled students
TMA1 (Tutor-Marked Assignment) ā€“ 1500 words
Write a report explaining the main accessibility challenges for disabled learners
that you work with or support in your own work context(s).
Critically evaluate the influence of the context (e.g. country, institution, perceived
role of online learning within education) on the: (1) identified challenges; (2)
influence of legislation; (3) roles and responsibilities of key individuals; (4) role of
assistive technologies in addressing these challenges.
TMA2 ā€“ 3000 words
Critically Evaluate your own learning resource in the following ways:
1. Briefly describe the resource and its accessibility features.
2. Evaluate the accessibility of your resource, identifying its strengths and
weaknesses.
3. Reflect on the processes of creating and evaluating accessible resources.
Context
ā€¢ Essays: Open University (UK) postgraduate assignments
ā€“ Distance learning, adult learners
ā€“ 1500+ words, free-text & open-ended questions
ā€¢ No ā€œGold Standardā€, wide range of content
ā€“ Perfect test ground for extractive techniques
ā€“ Impact of lack of (or limited) domain knowledge?
ā€¢ Bulk of activity (i.e. writing) takes place outside system
ā€“ Usage of drafts ā€œvaries a lotā€ among students
ā€“ Nature, scope and timing of feedback?
ā€¢ Limited possibility for ā€œmockā€ experiments:
ā€¢ testing & evaluation on ā€œliveā€ material
ā€¢ Connection with summative (tutor-based) assessment ?
openEssayist
localhost:8065
phaeros.open.ac.uk:80
openEssayist
PHP, Epiphany
[Symfony2]
User
openEssayist
RESTful API
PHP, Epiphany
User
User
pyEA
RESTful API
Python, Flask
localhost:8064
AfterTheDeadline
Spell/Grammar
checker
Java
User
localhost:9998
Apache Tika
Text Extractor
Java
Orchestrator
(Open)
Learner Model
pyEssayAnalyser
Python, NLTK
Extractive Summarisation
ā€¢ Hypothesis
ā€“ quality and position of key phrases and key sentences
within an essay (i.e., relative to the position of its
structural components) give idea of how complete and
well-structured the essay
ā€“ provide a basis for building suitable models of feedback
ā€¢ Experimenting with two simpler summarisation
strategies
ā€“ key phrase extraction : identifying individual words or
short phrases are the most suggestive of the content of a
discourse
ā€“ extractive summarisation: identifying whole key
sentences.
ā€¢ Rapid implementation and testing
Summarisation Processes
1. NL pre-processing of text
2. unsupervised recognition of structural
elements
3. unsupervised extraction of key
words/phrases
4. unsupervised extraction of key sentences.
Pre-processing
ā€¢ Using NLTK (Python-base Natural Language
Processing Toolkit)
ā€“ tokenisers,
ā€“ lemmatiser,
ā€“ part-of-speech tagger,
ā€“ List(s) of stop words.
ā€¢ Experimenting different approaches to define
suitable stop word list(s)
ā€“ domain-independent list?
ā€“ Generated from appropriate reference materials
(using TF-IDF, for example)?
Essay Structure
ā€¢ Restructure text as paragraphs/sentences
ā€¢ Automatic Identification of each paragraphā€™s structural role
ā€“ Summary, Introduction, conclusion, body, references, ā€¦
ā€“ Regardless of presence of content-specific headings
ā€“ No clues from formatting markup (plain text submission)
ā€¢ Decision trees developed through manual experimentation
ā€“ corpus of 135 student essays submitted in previous years for the
same module that the evaluation will be carried out on.
ā€¢ Still need formal evaluation but output good enough for
first rounds of OpenEssayist testing, and continually
improving
Key words, lemmas and phrases
ā€¢ Unsupervised extractive summarisation using graph-based ranking
methods (TextRank, Mihalcea & Tarau 2004, 2005)
ā€¢ Each unique word is represented by a node in the graph, and co-
occurrence relations (specifically, within-sentence word adjacency) are
represented by edges in the graph.
ā€¢ Compute a 'key-ness' value for each word in the essay ('Key-ness' can be
understood as 'significance within the context of the essayā€˜)
ā€¢ Centrality algorithm used to calculate the significance of each word
ā€“ betweenness centrality (Freeman 1977) and PageRank (Brin & Page 1998)
ā€“ Roughly speaking, a word with a high centrality score is a word that sits
adjacent to many other unique words which sit adjacent to many other unique
words whichā€¦, and so on.
ā€¢ The words with high(est) centrality scores are the key words.
ā€“ Decision needs to be made as to what proportion of the essay's words qualify
as key words.
ā€¢ Sequences of keywords in the surface text identify within-sentence key
phrases (bigrams, trigrams and quadgrams).
Key words, lemmas and phrases
Key Sentences
ā€¢ Similar graph-based ranking approach used to compute
key-ness scores for whole sentences.
ā€¢ Instead of word adjacency (as in the key word graph),
co-occurrence of words across pairs of sentences is the
relation used to construct the graph.
ā€“ similarity measures of every pair of sentences.
ā€¢ The similarity scores become edge weights in the
graph, while whole sentences become the nodes.
ā€¢ TextRank key sentence algorithm (based on PageRank
but with added edge weights) is then applied.
Extractive Summarisation - Sentences
Extractive Summarisation ā€“ Overview
Exploring The Design Space
ā¶Researcher-centred Design
ā€“ Data-driven
ā€“ Architecture setup, integration & refinement of
tools
ā€“ From discourse to summarisation
ā€“ Emerging properties, hypotheses building
ā€¢ Multiple External Representation
ā€¢ Mash-ups, reports, summaries, ā€¦
ā€¢ Highlighting co-occurrence of terms (or lack of)
ā€¢ Exploration & discovery, hypotheses building,
eliciting recommendations & heuristics
Exploring The Design Space
ā‘  Researcher-centred Design
ā€“ Data-driven
ā€“ Architecture setup, integration & refinement of tools
ā€“ From discourse to summarisation
ā€“ Emerging properties, hypotheses building
ā· Learner-centred Design
ā€“ Task-driven
ā€“ Hypotheses testing & validation, refinement
ā€“ From summarisation to formative feedback
ā€“ Live evaluation
Question: What kind of feedback?
Section of
essay
Purpose of section
Title
Write the full question (title) at the
top of your assignment. It will
contain keywords (known as
content and process words). See
the 'Understanding the question'
webpage for these.
Introduction
A paragraph or two to define key
terms and themes and indicate
how you intend to address the
question.
Main body
A series of paragraphs written in
full sentences that include specific
arguments relating to your answer.
Itā€™s vital to include evidence and
references to support your
arguments.
Conclusions
A short section to summarise main
points and findings. Try to focus on
the question but avoid repeating
what you wrote in the
introduction.
References
A list of sources (including module
materials) that are mentioned in
the essay.
ā€¢ Introductions
ā€“ An introduction provides your reader
with an overview of what your essay
will cover and what you want to say.
ā€“ Essays introductions should
ā€¢ set out the aims of the assignment and
signpost how your argument will unfold
ā€¢ introduce the issue and give any
essential background information
including a brief description of the
major debates that lie behind the
question
ā€¢ define the key words and terms
ā€¢ be between 5% and 10% of the total
word count
ā€“ Some students prefer to write the
introduction at an early stage, others
save it for when they have almost
completed the assignment. If you write
it early, don't allow it to constrain what
you want to write. It's a good idea to
check and revise the introduction after
the first draft.
ā€¢ The body of your essay
ā€“ ā€¦
Open University - Skills for OU Study
http://www.open.ac.uk/skillsforstudy/essays.php
Question: Reflective activities?
ā€¢ ā€œAdvice for actionā€
ā€“ Expectation vs. intention
ā€“ Reflection, self-report, validation of advices, ā€¦
ā€¢ Introducing user interventions in the system
ā€¢ Feeding back to the system? To the Essay Analyser?
Question: Drafts, History & Changes
Question: ā€œQualityā€ of output?
Current and Future Work
ā€¢ Three lines of experimentations:
ā€“ improve the different aspects of the essay analyser (e.g.
different ā€œkey-nessā€ metrics, introduce domain-specific
lists of stop-words)
ā€“ Analyses of summarisation output (e.g. factor analysis) to
run on existing corpus of essays
ā€¢ 5 years of essays on the H810 course, all marked and annotated by
human tutors),
ā€¢ identify trends and markers to be used as progress/performance
indicators;
ā€“ Iterative, user-centred, design and testing of openEssaysit
(refine possible usage scenarios, test pedagogical)
ā€¢ Currently proceeding with second design phase
ā€¢ First live evaluation, in authentic context, by a new
cohort of students on the H810 module (Sept 2013)

More Related Content

What's hot

Open IE tutorial 2018
Open IE tutorial 2018Open IE tutorial 2018
Open IE tutorial 2018Andre Freitas
Ā 
Assessment re-capitulated and questioned
Assessment re-capitulated and questionedAssessment re-capitulated and questioned
Assessment re-capitulated and questionedJames Atherton
Ā 
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate Writing
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate WritingLegitimising Knowersā€™ Multiple Voices in L2 Postgraduate Writing
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate WritingLok Ming Eric Cheung
Ā 
It services & research methods
It services & research methodsIt services & research methods
It services & research methodsAkanshShandilya
Ā 
A perspective on Conversational Agents
A perspective on Conversational AgentsA perspective on Conversational Agents
A perspective on Conversational AgentsZelia Blaga
Ā 
Ecology Through My Eyes
Ecology Through My EyesEcology Through My Eyes
Ecology Through My Eyeshksprague
Ā 
Effective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsEffective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsAndre Freitas
Ā 
ICBL DIscussion Forum
ICBL DIscussion ForumICBL DIscussion Forum
ICBL DIscussion Forummarorussell
Ā 
AI Beyond Deep Learning
AI Beyond Deep LearningAI Beyond Deep Learning
AI Beyond Deep LearningAndre Freitas
Ā 
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...Dilip Barad
Ā 
Tesol Presentation
Tesol PresentationTesol Presentation
Tesol Presentationcherepaha
Ā 

What's hot (12)

Open IE tutorial 2018
Open IE tutorial 2018Open IE tutorial 2018
Open IE tutorial 2018
Ā 
Assessment re-capitulated and questioned
Assessment re-capitulated and questionedAssessment re-capitulated and questioned
Assessment re-capitulated and questioned
Ā 
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate Writing
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate WritingLegitimising Knowersā€™ Multiple Voices in L2 Postgraduate Writing
Legitimising Knowersā€™ Multiple Voices in L2 Postgraduate Writing
Ā 
It services & research methods
It services & research methodsIt services & research methods
It services & research methods
Ā 
A perspective on Conversational Agents
A perspective on Conversational AgentsA perspective on Conversational Agents
A perspective on Conversational Agents
Ā 
Ecology Through My Eyes
Ecology Through My EyesEcology Through My Eyes
Ecology Through My Eyes
Ā 
Effective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP SystemsEffective Semantics for Engineering NLP Systems
Effective Semantics for Engineering NLP Systems
Ā 
ICBL DIscussion Forum
ICBL DIscussion ForumICBL DIscussion Forum
ICBL DIscussion Forum
Ā 
AI Beyond Deep Learning
AI Beyond Deep LearningAI Beyond Deep Learning
AI Beyond Deep Learning
Ā 
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...
Quantifying the Relation among Marks, Words and Time Duration in Descriptive ...
Ā 
Cv ning mei
Cv ning meiCv ning mei
Cv ning mei
Ā 
Tesol Presentation
Tesol PresentationTesol Presentation
Tesol Presentation
Ā 

Viewers also liked

Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...
Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...
Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...Poh-Sun Goh
Ā 
One (more) representation of the continuum of learning
One (more) representation of the continuum of learningOne (more) representation of the continuum of learning
One (more) representation of the continuum of learningPoh-Sun Goh
Ā 
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...Richard Oelmann
Ā 
Everything I have learnt about eLearning
Everything I have learnt about eLearningEverything I have learnt about eLearning
Everything I have learnt about eLearningPoh-Sun Goh
Ā 
Applied learning analytics to class and program learning activities
Applied learning analytics to class and program learning activitiesApplied learning analytics to class and program learning activities
Applied learning analytics to class and program learning activitiesPoh-Sun Goh
Ā 
The Value of Competency-based Medical Education Across the Continuum
The Value of Competency-based Medical Education Across the ContinuumThe Value of Competency-based Medical Education Across the Continuum
The Value of Competency-based Medical Education Across the ContinuumMedCouncilCan
Ā 
CBME and Assessment
CBME and Assessment CBME and Assessment
CBME and Assessment jakinyi
Ā 
Digital educational resource hyperlinked index vs restaurant menu
Digital educational resource hyperlinked index vs restaurant menuDigital educational resource hyperlinked index vs restaurant menu
Digital educational resource hyperlinked index vs restaurant menuPoh-Sun Goh
Ā 
SoLAR Storm talk: epistemology, pedagogy, assessment
SoLAR Storm talk: epistemology, pedagogy, assessmentSoLAR Storm talk: epistemology, pedagogy, assessment
SoLAR Storm talk: epistemology, pedagogy, assessmentSimon Knight
Ā 
Introduction to Competency-based Education
Introduction to Competency-based EducationIntroduction to Competency-based Education
Introduction to Competency-based EducationD2L
Ā 
Developing a multiple-document-processing performance assessment for epistem...
 Developing a multiple-document-processing performance assessment for epistem... Developing a multiple-document-processing performance assessment for epistem...
Developing a multiple-document-processing performance assessment for epistem...Simon Knight
Ā 
Using data (analytics:analysis) to guide (e)teaching and (e)learning
Using data (analytics:analysis) to guide (e)teaching and (e)learningUsing data (analytics:analysis) to guide (e)teaching and (e)learning
Using data (analytics:analysis) to guide (e)teaching and (e)learningPoh-Sun Goh
Ā 
Using data to guide (e)teaching and (e)learning
Using data to guide (e)teaching and (e)learningUsing data to guide (e)teaching and (e)learning
Using data to guide (e)teaching and (e)learningPoh-Sun Goh
Ā 
Personal and personalised learning and teaching
Personal and personalised learning and teachingPersonal and personalised learning and teaching
Personal and personalised learning and teachingPoh-Sun Goh
Ā 
Application of learning analytics to case studies
Application of learning analytics to case studiesApplication of learning analytics to case studies
Application of learning analytics to case studiesPoh-Sun Goh
Ā 
The continuum of learning
The continuum of learningThe continuum of learning
The continuum of learningPoh-Sun Goh
Ā 
edmedia2014-learning-analytics-keynote
edmedia2014-learning-analytics-keynoteedmedia2014-learning-analytics-keynote
edmedia2014-learning-analytics-keynoteSimon Buckingham Shum
Ā 
Learning Analytics - George Siemens
Learning Analytics - George SiemensLearning Analytics - George Siemens
Learning Analytics - George SiemensSungjin Nam
Ā 
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...Anna De Liddo
Ā 
Lak 05
Lak 05Lak 05
Lak 05dws1d
Ā 

Viewers also liked (20)

Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...
Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...
Reflect on how Bloom's taxonomy, Miller's pyramid and the Kirkpatrick model m...
Ā 
One (more) representation of the continuum of learning
One (more) representation of the continuum of learningOne (more) representation of the continuum of learning
One (more) representation of the continuum of learning
Ā 
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...
Learning Outcomes and Assessment - Achieving Constructive Alignment Treforest...
Ā 
Everything I have learnt about eLearning
Everything I have learnt about eLearningEverything I have learnt about eLearning
Everything I have learnt about eLearning
Ā 
Applied learning analytics to class and program learning activities
Applied learning analytics to class and program learning activitiesApplied learning analytics to class and program learning activities
Applied learning analytics to class and program learning activities
Ā 
The Value of Competency-based Medical Education Across the Continuum
The Value of Competency-based Medical Education Across the ContinuumThe Value of Competency-based Medical Education Across the Continuum
The Value of Competency-based Medical Education Across the Continuum
Ā 
CBME and Assessment
CBME and Assessment CBME and Assessment
CBME and Assessment
Ā 
Digital educational resource hyperlinked index vs restaurant menu
Digital educational resource hyperlinked index vs restaurant menuDigital educational resource hyperlinked index vs restaurant menu
Digital educational resource hyperlinked index vs restaurant menu
Ā 
SoLAR Storm talk: epistemology, pedagogy, assessment
SoLAR Storm talk: epistemology, pedagogy, assessmentSoLAR Storm talk: epistemology, pedagogy, assessment
SoLAR Storm talk: epistemology, pedagogy, assessment
Ā 
Introduction to Competency-based Education
Introduction to Competency-based EducationIntroduction to Competency-based Education
Introduction to Competency-based Education
Ā 
Developing a multiple-document-processing performance assessment for epistem...
 Developing a multiple-document-processing performance assessment for epistem... Developing a multiple-document-processing performance assessment for epistem...
Developing a multiple-document-processing performance assessment for epistem...
Ā 
Using data (analytics:analysis) to guide (e)teaching and (e)learning
Using data (analytics:analysis) to guide (e)teaching and (e)learningUsing data (analytics:analysis) to guide (e)teaching and (e)learning
Using data (analytics:analysis) to guide (e)teaching and (e)learning
Ā 
Using data to guide (e)teaching and (e)learning
Using data to guide (e)teaching and (e)learningUsing data to guide (e)teaching and (e)learning
Using data to guide (e)teaching and (e)learning
Ā 
Personal and personalised learning and teaching
Personal and personalised learning and teachingPersonal and personalised learning and teaching
Personal and personalised learning and teaching
Ā 
Application of learning analytics to case studies
Application of learning analytics to case studiesApplication of learning analytics to case studies
Application of learning analytics to case studies
Ā 
The continuum of learning
The continuum of learningThe continuum of learning
The continuum of learning
Ā 
edmedia2014-learning-analytics-keynote
edmedia2014-learning-analytics-keynoteedmedia2014-learning-analytics-keynote
edmedia2014-learning-analytics-keynote
Ā 
Learning Analytics - George Siemens
Learning Analytics - George SiemensLearning Analytics - George Siemens
Learning Analytics - George Siemens
Ā 
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...
The Evidence Hub: Harnessing the Collective Intelligence of Communities to Bu...
Ā 
Lak 05
Lak 05Lak 05
Lak 05
Ā 

Similar to OpenEssayist: Extractive Summarisation and Formative Assessment (DCLA13)

2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky
2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky
2016-05-31 Venia Legendi (CEITER): Sergey Sosnovskyifi8106tlu
Ā 
Systematic Literature Reviews : Concise Overview
Systematic Literature Reviews : Concise OverviewSystematic Literature Reviews : Concise Overview
Systematic Literature Reviews : Concise Overviewyoukayaslam
Ā 
Publication workshop
Publication workshopPublication workshop
Publication workshopFrances Bell
Ā 
empirical-SLR.pptx
empirical-SLR.pptxempirical-SLR.pptx
empirical-SLR.pptxJitha Kannan
Ā 
La notes (5 10)
La notes (5 10)La notes (5 10)
La notes (5 10)hakim azman
Ā 
eAssessment in practice
eAssessment in practiceeAssessment in practice
eAssessment in practiceIan Brown
Ā 
Near Real-time Web-Page Recs Using Content Features
Near Real-time Web-Page Recs Using Content FeaturesNear Real-time Web-Page Recs Using Content Features
Near Real-time Web-Page Recs Using Content FeaturesAshok Venkatesan
Ā 
Systematic Literature Reviews and Systematic Mapping Studies
Systematic Literature Reviews and Systematic Mapping StudiesSystematic Literature Reviews and Systematic Mapping Studies
Systematic Literature Reviews and Systematic Mapping Studiesalessio_ferrari
Ā 
Data Acquisition for Sentiment Analysis
Data Acquisition for Sentiment AnalysisData Acquisition for Sentiment Analysis
Data Acquisition for Sentiment AnalysisAli BELCAID
Ā 
Introduction & Literature Review Webinar
Introduction & Literature Review WebinarIntroduction & Literature Review Webinar
Introduction & Literature Review WebinarStatistics Solutions
Ā 
Presentation - Systematic Review - March23.pptx
Presentation - Systematic Review - March23.pptxPresentation - Systematic Review - March23.pptx
Presentation - Systematic Review - March23.pptxChetan Acharya
Ā 
Final presentation
Final presentationFinal presentation
Final presentationNitish Upreti
Ā 
Knowledge Representation on the Web
Knowledge Representation on the WebKnowledge Representation on the Web
Knowledge Representation on the WebRinke Hoekstra
Ā 
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRR
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRRr_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRR
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRRMarvinCastaneda7
Ā 
Research Writing - Universitas Indonesia
Research Writing - Universitas IndonesiaResearch Writing - Universitas Indonesia
Research Writing - Universitas IndonesiaFariz Darari
Ā 
Intro to Deep Learning for Question Answering
Intro to Deep Learning for Question AnsweringIntro to Deep Learning for Question Answering
Intro to Deep Learning for Question AnsweringTraian Rebedea
Ā 
Hackathon report catalogue-ontology-vocabulary-characteristcs-relevant-to-e...
Hackathon report   catalogue-ontology-vocabulary-characteristcs-relevant-to-e...Hackathon report   catalogue-ontology-vocabulary-characteristcs-relevant-to-e...
Hackathon report catalogue-ontology-vocabulary-characteristcs-relevant-to-e...Amanda Vizedom
Ā 

Similar to OpenEssayist: Extractive Summarisation and Formative Assessment (DCLA13) (20)

2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky
2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky
2016-05-31 Venia Legendi (CEITER): Sergey Sosnovsky
Ā 
Systematic Literature Reviews : Concise Overview
Systematic Literature Reviews : Concise OverviewSystematic Literature Reviews : Concise Overview
Systematic Literature Reviews : Concise Overview
Ā 
Publication workshop
Publication workshopPublication workshop
Publication workshop
Ā 
empirical-SLR.pptx
empirical-SLR.pptxempirical-SLR.pptx
empirical-SLR.pptx
Ā 
La notes (5 10)
La notes (5 10)La notes (5 10)
La notes (5 10)
Ā 
eAssessment in practice
eAssessment in practiceeAssessment in practice
eAssessment in practice
Ā 
Near Real-time Web-Page Recs Using Content Features
Near Real-time Web-Page Recs Using Content FeaturesNear Real-time Web-Page Recs Using Content Features
Near Real-time Web-Page Recs Using Content Features
Ā 
Systematic Literature Reviews and Systematic Mapping Studies
Systematic Literature Reviews and Systematic Mapping StudiesSystematic Literature Reviews and Systematic Mapping Studies
Systematic Literature Reviews and Systematic Mapping Studies
Ā 
Data Acquisition for Sentiment Analysis
Data Acquisition for Sentiment AnalysisData Acquisition for Sentiment Analysis
Data Acquisition for Sentiment Analysis
Ā 
Introduction & Literature Review Webinar
Introduction & Literature Review WebinarIntroduction & Literature Review Webinar
Introduction & Literature Review Webinar
Ā 
R vaughan
R vaughanR vaughan
R vaughan
Ā 
Designing e-Learning Objects
Designing e-Learning ObjectsDesigning e-Learning Objects
Designing e-Learning Objects
Ā 
Presentation - Systematic Review - March23.pptx
Presentation - Systematic Review - March23.pptxPresentation - Systematic Review - March23.pptx
Presentation - Systematic Review - March23.pptx
Ā 
Final presentation
Final presentationFinal presentation
Final presentation
Ā 
Knowledge Representation on the Web
Knowledge Representation on the WebKnowledge Representation on the Web
Knowledge Representation on the Web
Ā 
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRR
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRRr_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRR
r_vaughan.pptRRRRRRRRRRRRRRRRRRRRRRRRRRRR
Ā 
Research Writing - Universitas Indonesia
Research Writing - Universitas IndonesiaResearch Writing - Universitas Indonesia
Research Writing - Universitas Indonesia
Ā 
Intro to Deep Learning for Question Answering
Intro to Deep Learning for Question AnsweringIntro to Deep Learning for Question Answering
Intro to Deep Learning for Question Answering
Ā 
Hackathon report catalogue-ontology-vocabulary-characteristcs-relevant-to-e...
Hackathon report   catalogue-ontology-vocabulary-characteristcs-relevant-to-e...Hackathon report   catalogue-ontology-vocabulary-characteristcs-relevant-to-e...
Hackathon report catalogue-ontology-vocabulary-characteristcs-relevant-to-e...
Ā 
MASHLM 1st year thesis workshop
MASHLM 1st year thesis workshopMASHLM 1st year thesis workshop
MASHLM 1st year thesis workshop
Ā 

More from Nicolas Van Labeke

Personalised and Adaptive Mentoring in Medical Education - The myPAL project
Personalised and Adaptive Mentoring in Medical Education - The myPAL projectPersonalised and Adaptive Mentoring in Medical Education - The myPAL project
Personalised and Adaptive Mentoring in Medical Education - The myPAL projectNicolas Van Labeke
Ā 
Auditory perceptual learning and game design an example for tinnitus manage...
Auditory perceptual learning and game design   an example for tinnitus manage...Auditory perceptual learning and game design   an example for tinnitus manage...
Auditory perceptual learning and game design an example for tinnitus manage...Nicolas Van Labeke
Ā 
MyPlan - similarity metrics for matching lifelong learner timelines
MyPlan - similarity metrics for matching lifelong learner timelinesMyPlan - similarity metrics for matching lifelong learner timelines
MyPlan - similarity metrics for matching lifelong learner timelinesNicolas Van Labeke
Ā 
Principled design of game-based auditory learning environments: Examples of A...
Principled design of game-based auditory learning environments: Examples of A...Principled design of game-based auditory learning environments: Examples of A...
Principled design of game-based auditory learning environments: Examples of A...Nicolas Van Labeke
Ā 
Using Similarity Metrics for Matching Lifelong Learners
Using Similarity Metrics for Matching Lifelong LearnersUsing Similarity Metrics for Matching Lifelong Learners
Using Similarity Metrics for Matching Lifelong LearnersNicolas Van Labeke
Ā 
Intrinsic Integration and the Design of Games for Auditory Perceptual Learning
Intrinsic Integration and the Design of Games for Auditory Perceptual LearningIntrinsic Integration and the Design of Games for Auditory Perceptual Learning
Intrinsic Integration and the Design of Games for Auditory Perceptual LearningNicolas Van Labeke
Ā 

More from Nicolas Van Labeke (6)

Personalised and Adaptive Mentoring in Medical Education - The myPAL project
Personalised and Adaptive Mentoring in Medical Education - The myPAL projectPersonalised and Adaptive Mentoring in Medical Education - The myPAL project
Personalised and Adaptive Mentoring in Medical Education - The myPAL project
Ā 
Auditory perceptual learning and game design an example for tinnitus manage...
Auditory perceptual learning and game design   an example for tinnitus manage...Auditory perceptual learning and game design   an example for tinnitus manage...
Auditory perceptual learning and game design an example for tinnitus manage...
Ā 
MyPlan - similarity metrics for matching lifelong learner timelines
MyPlan - similarity metrics for matching lifelong learner timelinesMyPlan - similarity metrics for matching lifelong learner timelines
MyPlan - similarity metrics for matching lifelong learner timelines
Ā 
Principled design of game-based auditory learning environments: Examples of A...
Principled design of game-based auditory learning environments: Examples of A...Principled design of game-based auditory learning environments: Examples of A...
Principled design of game-based auditory learning environments: Examples of A...
Ā 
Using Similarity Metrics for Matching Lifelong Learners
Using Similarity Metrics for Matching Lifelong LearnersUsing Similarity Metrics for Matching Lifelong Learners
Using Similarity Metrics for Matching Lifelong Learners
Ā 
Intrinsic Integration and the Design of Games for Auditory Perceptual Learning
Intrinsic Integration and the Design of Games for Auditory Perceptual LearningIntrinsic Integration and the Design of Games for Auditory Perceptual Learning
Intrinsic Integration and the Design of Games for Auditory Perceptual Learning
Ā 

Recently uploaded

Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024The Digital Insurer
Ā 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProduct Anonymous
Ā 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024The Digital Insurer
Ā 
šŸ¬ The future of MySQL is Postgres šŸ˜
šŸ¬  The future of MySQL is Postgres   šŸ˜šŸ¬  The future of MySQL is Postgres   šŸ˜
šŸ¬ The future of MySQL is Postgres šŸ˜RTylerCroy
Ā 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsJoaquim Jorge
Ā 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024The Digital Insurer
Ā 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...Neo4j
Ā 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
Ā 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024The Digital Insurer
Ā 
Top 5 Benefits OF Using Muvi Live Paywall For Live Streams
Top 5 Benefits OF Using Muvi Live Paywall For Live StreamsTop 5 Benefits OF Using Muvi Live Paywall For Live Streams
Top 5 Benefits OF Using Muvi Live Paywall For Live StreamsRoshan Dwivedi
Ā 
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone ProcessorsExploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processorsdebabhi2
Ā 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businesspanagenda
Ā 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CVKhem
Ā 
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...Principled Technologies
Ā 
AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAndrey Devyatkin
Ā 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024The Digital Insurer
Ā 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FMESafe Software
Ā 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherRemote DBA Services
Ā 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
Ā 

Recently uploaded (20)

+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
+971581248768>> SAFE AND ORIGINAL ABORTION PILLS FOR SALE IN DUBAI AND ABUDHA...
Ā 
Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024Manulife - Insurer Innovation Award 2024
Manulife - Insurer Innovation Award 2024
Ā 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
Ā 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Ā 
šŸ¬ The future of MySQL is Postgres šŸ˜
šŸ¬  The future of MySQL is Postgres   šŸ˜šŸ¬  The future of MySQL is Postgres   šŸ˜
šŸ¬ The future of MySQL is Postgres šŸ˜
Ā 
Artificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and MythsArtificial Intelligence: Facts and Myths
Artificial Intelligence: Facts and Myths
Ā 
Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024Partners Life - Insurer Innovation Award 2024
Partners Life - Insurer Innovation Award 2024
Ā 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Ā 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
Ā 
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024Tata AIG General Insurance Company - Insurer Innovation Award 2024
Tata AIG General Insurance Company - Insurer Innovation Award 2024
Ā 
Top 5 Benefits OF Using Muvi Live Paywall For Live Streams
Top 5 Benefits OF Using Muvi Live Paywall For Live StreamsTop 5 Benefits OF Using Muvi Live Paywall For Live Streams
Top 5 Benefits OF Using Muvi Live Paywall For Live Streams
Ā 
Exploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone ProcessorsExploring the Future Potential of AI-Enabled Smartphone Processors
Exploring the Future Potential of AI-Enabled Smartphone Processors
Ā 
Why Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire businessWhy Teams call analytics are critical to your entire business
Why Teams call analytics are critical to your entire business
Ā 
Real Time Object Detection Using Open CV
Real Time Object Detection Using Open CVReal Time Object Detection Using Open CV
Real Time Object Detection Using Open CV
Ā 
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...
Deploy with confidence: VMware Cloud Foundation 5.1 on next gen Dell PowerEdg...
Ā 
AWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of TerraformAWS Community Day CPH - Three problems of Terraform
AWS Community Day CPH - Three problems of Terraform
Ā 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Ā 
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers:  A Deep Dive into Serverless Spatial Data and FMECloud Frontiers:  A Deep Dive into Serverless Spatial Data and FME
Cloud Frontiers: A Deep Dive into Serverless Spatial Data and FME
Ā 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
Ā 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
Ā 

OpenEssayist: Extractive Summarisation and Formative Assessment (DCLA13)

  • 1. OpenEssayist: Extractive Summarisation and Formative Assessment of Free-Text Essays Nicolas Van Labekeļ‚, Denise Whitelock ļ‚, Debora Field ļ‚‚, Stephen Pulmanļ‚‚, John Richardson ļ‚ ļ‚ Institute of Educational Technology ā€“ The Open University ļ‚‚ Department of Computer Science ā€“ University of Oxford
  • 2. SAFeSEA: Research Questions ā€¢ How can an automated system detect passages on which a human marker would usually give some feedback ? ā€¢ Can existing methods of information extraction, summarization be adapted to select content for such feedback ? ā€¢ How effectively can these methods deliver feedback ? ā€¢ What effect does these techniques have on essay improvement? On current essay and in future ones ? On self-regulation and metacognition ?
  • 3. Context ā€¢ Essays: Open University (UK) postgraduate assignments ā€“ Distance learning, adult learners ā€“ 1500+ words, free-text & open-ended questions ā€¢ No ā€œGold Standardā€, wide range of content ā€“ Perfect test ground for extractive techniques ā€“ Impact of lack of (or limited) domain knowledge? ā€¢ Bulk of activity (i.e. writing) takes place outside system ā€“ Usage of drafts ā€œvaries a lotā€ among students ā€“ Nature, scope and timing of feedback? ā€¢ Limited possibility for ā€œmockā€ experiments: ā€¢ testing & evaluation on ā€œliveā€ material ā€¢ Connection with summative (tutor-based) assessment ?
  • 4. Education Postgraduate Course H810 Accessible online learning: supporting disabled students TMA1 (Tutor-Marked Assignment) ā€“ 1500 words Write a report explaining the main accessibility challenges for disabled learners that you work with or support in your own work context(s). Critically evaluate the influence of the context (e.g. country, institution, perceived role of online learning within education) on the: (1) identified challenges; (2) influence of legislation; (3) roles and responsibilities of key individuals; (4) role of assistive technologies in addressing these challenges. TMA2 ā€“ 3000 words Critically Evaluate your own learning resource in the following ways: 1. Briefly describe the resource and its accessibility features. 2. Evaluate the accessibility of your resource, identifying its strengths and weaknesses. 3. Reflect on the processes of creating and evaluating accessible resources.
  • 5. Context ā€¢ Essays: Open University (UK) postgraduate assignments ā€“ Distance learning, adult learners ā€“ 1500+ words, free-text & open-ended questions ā€¢ No ā€œGold Standardā€, wide range of content ā€“ Perfect test ground for extractive techniques ā€“ Impact of lack of (or limited) domain knowledge? ā€¢ Bulk of activity (i.e. writing) takes place outside system ā€“ Usage of drafts ā€œvaries a lotā€ among students ā€“ Nature, scope and timing of feedback? ā€¢ Limited possibility for ā€œmockā€ experiments: ā€¢ testing & evaluation on ā€œliveā€ material ā€¢ Connection with summative (tutor-based) assessment ?
  • 6. openEssayist localhost:8065 phaeros.open.ac.uk:80 openEssayist PHP, Epiphany [Symfony2] User openEssayist RESTful API PHP, Epiphany User User pyEA RESTful API Python, Flask localhost:8064 AfterTheDeadline Spell/Grammar checker Java User localhost:9998 Apache Tika Text Extractor Java Orchestrator (Open) Learner Model pyEssayAnalyser Python, NLTK
  • 7. Extractive Summarisation ā€¢ Hypothesis ā€“ quality and position of key phrases and key sentences within an essay (i.e., relative to the position of its structural components) give idea of how complete and well-structured the essay ā€“ provide a basis for building suitable models of feedback ā€¢ Experimenting with two simpler summarisation strategies ā€“ key phrase extraction : identifying individual words or short phrases are the most suggestive of the content of a discourse ā€“ extractive summarisation: identifying whole key sentences. ā€¢ Rapid implementation and testing
  • 8. Summarisation Processes 1. NL pre-processing of text 2. unsupervised recognition of structural elements 3. unsupervised extraction of key words/phrases 4. unsupervised extraction of key sentences.
  • 9. Pre-processing ā€¢ Using NLTK (Python-base Natural Language Processing Toolkit) ā€“ tokenisers, ā€“ lemmatiser, ā€“ part-of-speech tagger, ā€“ List(s) of stop words. ā€¢ Experimenting different approaches to define suitable stop word list(s) ā€“ domain-independent list? ā€“ Generated from appropriate reference materials (using TF-IDF, for example)?
  • 10. Essay Structure ā€¢ Restructure text as paragraphs/sentences ā€¢ Automatic Identification of each paragraphā€™s structural role ā€“ Summary, Introduction, conclusion, body, references, ā€¦ ā€“ Regardless of presence of content-specific headings ā€“ No clues from formatting markup (plain text submission) ā€¢ Decision trees developed through manual experimentation ā€“ corpus of 135 student essays submitted in previous years for the same module that the evaluation will be carried out on. ā€¢ Still need formal evaluation but output good enough for first rounds of OpenEssayist testing, and continually improving
  • 11.
  • 12. Key words, lemmas and phrases ā€¢ Unsupervised extractive summarisation using graph-based ranking methods (TextRank, Mihalcea & Tarau 2004, 2005) ā€¢ Each unique word is represented by a node in the graph, and co- occurrence relations (specifically, within-sentence word adjacency) are represented by edges in the graph. ā€¢ Compute a 'key-ness' value for each word in the essay ('Key-ness' can be understood as 'significance within the context of the essayā€˜) ā€¢ Centrality algorithm used to calculate the significance of each word ā€“ betweenness centrality (Freeman 1977) and PageRank (Brin & Page 1998) ā€“ Roughly speaking, a word with a high centrality score is a word that sits adjacent to many other unique words which sit adjacent to many other unique words whichā€¦, and so on. ā€¢ The words with high(est) centrality scores are the key words. ā€“ Decision needs to be made as to what proportion of the essay's words qualify as key words. ā€¢ Sequences of keywords in the surface text identify within-sentence key phrases (bigrams, trigrams and quadgrams).
  • 13. Key words, lemmas and phrases
  • 14. Key Sentences ā€¢ Similar graph-based ranking approach used to compute key-ness scores for whole sentences. ā€¢ Instead of word adjacency (as in the key word graph), co-occurrence of words across pairs of sentences is the relation used to construct the graph. ā€“ similarity measures of every pair of sentences. ā€¢ The similarity scores become edge weights in the graph, while whole sentences become the nodes. ā€¢ TextRank key sentence algorithm (based on PageRank but with added edge weights) is then applied.
  • 17. Exploring The Design Space ā¶Researcher-centred Design ā€“ Data-driven ā€“ Architecture setup, integration & refinement of tools ā€“ From discourse to summarisation ā€“ Emerging properties, hypotheses building
  • 18.
  • 19.
  • 20.
  • 21.
  • 22.
  • 23.
  • 24.
  • 25. ā€¢ Multiple External Representation ā€¢ Mash-ups, reports, summaries, ā€¦ ā€¢ Highlighting co-occurrence of terms (or lack of) ā€¢ Exploration & discovery, hypotheses building, eliciting recommendations & heuristics
  • 26. Exploring The Design Space ā‘  Researcher-centred Design ā€“ Data-driven ā€“ Architecture setup, integration & refinement of tools ā€“ From discourse to summarisation ā€“ Emerging properties, hypotheses building ā· Learner-centred Design ā€“ Task-driven ā€“ Hypotheses testing & validation, refinement ā€“ From summarisation to formative feedback ā€“ Live evaluation
  • 27. Question: What kind of feedback?
  • 28. Section of essay Purpose of section Title Write the full question (title) at the top of your assignment. It will contain keywords (known as content and process words). See the 'Understanding the question' webpage for these. Introduction A paragraph or two to define key terms and themes and indicate how you intend to address the question. Main body A series of paragraphs written in full sentences that include specific arguments relating to your answer. Itā€™s vital to include evidence and references to support your arguments. Conclusions A short section to summarise main points and findings. Try to focus on the question but avoid repeating what you wrote in the introduction. References A list of sources (including module materials) that are mentioned in the essay. ā€¢ Introductions ā€“ An introduction provides your reader with an overview of what your essay will cover and what you want to say. ā€“ Essays introductions should ā€¢ set out the aims of the assignment and signpost how your argument will unfold ā€¢ introduce the issue and give any essential background information including a brief description of the major debates that lie behind the question ā€¢ define the key words and terms ā€¢ be between 5% and 10% of the total word count ā€“ Some students prefer to write the introduction at an early stage, others save it for when they have almost completed the assignment. If you write it early, don't allow it to constrain what you want to write. It's a good idea to check and revise the introduction after the first draft. ā€¢ The body of your essay ā€“ ā€¦ Open University - Skills for OU Study http://www.open.ac.uk/skillsforstudy/essays.php
  • 29. Question: Reflective activities? ā€¢ ā€œAdvice for actionā€ ā€“ Expectation vs. intention ā€“ Reflection, self-report, validation of advices, ā€¦ ā€¢ Introducing user interventions in the system ā€¢ Feeding back to the system? To the Essay Analyser?
  • 32. Current and Future Work ā€¢ Three lines of experimentations: ā€“ improve the different aspects of the essay analyser (e.g. different ā€œkey-nessā€ metrics, introduce domain-specific lists of stop-words) ā€“ Analyses of summarisation output (e.g. factor analysis) to run on existing corpus of essays ā€¢ 5 years of essays on the H810 course, all marked and annotated by human tutors), ā€¢ identify trends and markers to be used as progress/performance indicators; ā€“ Iterative, user-centred, design and testing of openEssaysit (refine possible usage scenarios, test pedagogical) ā€¢ Currently proceeding with second design phase ā€¢ First live evaluation, in authentic context, by a new cohort of students on the H810 module (Sept 2013)

Editor's Notes

  1. Before extracting key terms and sentences from the text, the text is automatically pre-processed using some modules from the Natural Language Processing Toolkit (Bird et al. 2009): several tokenisers, a lemmatiser, a part-of-speech tagger, and a list of stop words. We are experimenting with different approaches to defining a suitable stop word list, and are not yet decided whether to use a domain-independent list or whether to use a domain-specific list derived from appropriate reference materials (using TF-IDF, for example).Lemmatization usually refers to doing things properly with the use of a vocabulary and morphological analysis of words, normally aiming to remove inflectional endings only and to return the base or dictionary form of a word, which is known as the lemma
  2. The identification of the essay structure is carried out using decision trees developed through manual experimentation with a corpus of 135 student essays submitted in previous years for the same module that the evaluation will be carried out on. The system automatically recognises which structural role is played by each paragraph in the essay (including summary, introduction, conclusion, main body, references, etc.). This identification is achieved regardless of the presence of content-specific headings and without getting clues from formatting mark-up. We have not yet carried out a formal evaluation of the structure identification procedure, but its accuracy rates are good enough to use in first rounds of OpenEssayist testing, and are continually improving.
  3. Essay Analyser uses graph-based ranking methods to perform unsupervised extractive summarisation, following TextRank (Mihalcea & Tarau 2004, 2005). One graph is used to derive key words and short phrases, and a second graph is used for the derivation of key sentences. Regarding key words, to compute a 'key-ness' value for each word in the essay, each unique word is represented by a node in the graph, and co-occurrence relations (specifically, within-sentence word adjacency) are represented by edges in the graph. 'Key-ness' can be understood as 'significance within the context of the essay'. A centrality algorithm ā€“ we have experimented with betweenness centrality(Freeman 1977) and PageRank (Brin & Page 1998) ā€“ is used to calculate the significance of each word. Roughly speaking, a word with a high centrality score is a word that sits adjacent to many other unique words which sit adjacent to many other unique words whichā€¦, and so on. The words with high centrality scores are the key words. Since a centrality score is attributed to every unique word in the essay, a decision needs to be made as to what proportion of the essay's words qualify as key words. The key word distribution of scores follows the same shape for all essays, an acute elbow and then a very long tail, observed for word adjacency graphs by (FerreriCancho & SolĆ© 2001). We therefore currently take the key-ness threshold to be the place where the elbow bend appears by eye to be sharpest. We are investigating alternative and less subjective methods of deciding where the threshold should be (e.g., investigating graph structure through randomisation methods). Once key words have been identified, the system matches sequences of these against the surface text to identify within-sentence key phrases (bigrams, trigrams and quadgrams).In fact the graph nodes are the lemmas of the unique words, but for brevity's sake, we will speak in terms of words.
  4. A similar graph-based ranking approach is used to compute key-ness scores to rank the essay's sentences. Instead of word adjacency (as in the key word graph), co-occurrence of words across pairs of sentences is the relation used to construct the graph. More specifically, we currently use cosine similarity to derive a similarity score for every pair of sentences. The similarity scores become edge weights in the graph, while whole sentences become the nodes. The TextRank key sentence algorithm (based on PageRank but with added edge weights) is then applied. We are intending to experiment with alternative similarity measures, including vector space measures of word similarity originally described in (SchĆ¼tze 1998).
  5. Vagaries : An unexpected and inexplicable change in something (in a situation or a person's behaviour, etc.)Rule-of Thumb: A rule or principle that provides guidance to appropriate behaviour
  6. Vagaries : An unexpected and inexplicable change in something (in a situation or a person's behaviour, etc.)Rule-of Thumb: A rule or principle that provides guidance to appropriate behaviour