Presentation given to the London Argumentation Forum on 2012-04-20, based on work by Jodi Schneider, Adam Wyner, Katie Atkinson, Trevor Bench-Capon. Closely related to our COMMA 2012 presentation.
Event homepage for LAF 2012: http://www.dcs.kcl.ac.uk/pg/hadjinik/LAF/programme.html
For details see our COMMA paper: http://jodischneider.com/pubs/comma2012.pdf
Identifying arguments for evaluation using an argument explorer - London Argumentation Forum - 2012-04-20
1. Identifying Arguments for Evaluation
using an Argument Explorer
Jodi Schneider1, Adam Wyner2, Katie Atkinson2, Trevor Bench-Capon2
1DigitalEnterprise Research Institute, National University of Ireland
2Department of Computer Science, University of Liverpool
April 20, 2012
London Argumentation Forum
6. Goals
• Extract arguments from source texts so they
can be evaluated with formal automated tools
• Speed the work of human analysts
• Make argument identification more objective
April 20, 2012 London Argumentation Forum
7. Strategy & Issues
• Decompose the complexity of a text
– What are the parts of an argument?
– What kind of domain knowledge do we need?
– How are the parts of the argument related?
– What are the contrasts and negations from which
we can derive attack relationships?
April 20, 2012 London Argumentation Forum
8. Use case:
Which camera should I buy?
April 20, 2012 London Argumentation Forum
9. Value-based Practical Reasoning
Argumentation Scheme
Premises:
Before doing action A, the current circumstances are R;
After doing action A, the new circumstances are S;
G is a goal of the agent Ag, where S implies G;
Doing action A in R and achieving G promotes value V;
Conclusion:
We should perform action A.
April 20, 2012 London Argumentation Forum
10. Consumer Argumentation Scheme
Premises:
Camera X has property P.
Property P promotes value V for agent A.
Conclusion:
Agent A should Action1 Camera X.
April 20, 2012 London Argumentation Forum
11. Critical Questions
• Does Camera X have property P?
• Does property P promote value V for agent A?
• Is value V more important than value V’ for
agent A?
April 20, 2012 London Argumentation Forum
12. Analyst’s goal: instantiate
Premises:
The Canon SX220 has good video quality.
Good video quality promotes image quality for
casual photographers.
Conclusion:
Casual photographers should buy the Canon
SX220.
April 20, 2012 London Argumentation Forum
13. … starting from this
April 20, 2012 London Argumentation Forum
14. Highlight parts of the argument
• Does Camera X have property P?
• Does property P promote value V for agent A?
• Is value V more important than value V’ for
agent A?
April 20, 2012 London Argumentation Forum
15. Highlight parts of the argument
• Argumentative indicators
• Property – with camera terminology
• Value for agent – with sentiment, user models
• Value V more important – with comparisons
April 20, 2012 London Argumentation Forum
16. Implementing with a Text Analysis Tool
• GATE “General Architecture for Text Engineering”
• Environment for text analysis
• Adds annotation to text
– Highlight annotations with
– Search for annotations
– Can work with large corpora of text
– Coarse or fine-grained annotations
April 20, 2012 London Argumentation Forum
17. Help analysts find relevant passages
• Indicators of
after, as, because, for, since, when, ....
• Indicators of
therefore, in conclusion, consequently, ....
• Indicators of contrast
but, except, not, never, no, ....
April 20, 2012 London Argumentation Forum
23. Agents: User Models
• User’s parameters
Age, gender, education, previous camera experience, ....
• User’s context of use
Party, indoors, sport, travel, desired output format, ....
• User’s constraints
Cost, portability, size, richness or flexibility of features, ....
• User’s quality expectations
Colour quality, information density, reliability, ....
April 20, 2012 London Argumentation Forum
24. Instantiating the CAS
Premises:
The Canon SX220 camera has property P.
Property P promotes value V for agent A.
Conclusion:
Agent A should buy the Canon SX220.
April 20, 2012 London Argumentation Forum
27. An argument for buying the camera
Premises:
The pictures are perfectly exposed.
The pictures are well-focused.
No camera shake.
Good video quality.
Each of these properties promotes image quality.
Conclusion:
(You, the reader,) should buy the CanonSX220.
April 20, 2012 London Argumentation Forum
28. An argument for NOT buying the
camera
Premises:
The colour is poor when using the flash.
The images are not crisp when using the flash.
The flash causes a shadow.
Each of these properties demotes image quality.
Conclusion:
(You, the reader,) should NOT buy the CanonSX220.
April 20, 2012 London Argumentation Forum
29. Counterarguments to the premises of
“Don’t buy”
The colour is poor when using the flash.
For good colour, use the colour setting, not the
flash.
The images are not crisp when using the flash.
No need to use flash even in low light.
The flash causes a shadow.
There is a corrective video about the flash shadow.
April 20, 2012 London Argumentation Forum
30. Future Work
• Tool refinement
• Add terminology modules to the tool
• User models – how do they play a role
• More complicated query patterns, what results
do we get?
• More elaborate examples
• Disambiguation issues for rhetorical terminology
– must deal with it step-by-step, what are the
indicators we can use to disambiguate
April 20, 2012 London Argumentation Forum
31. Thanks to our funders!
• FP7-ICT-2009-4 Programme, IMPACT Project,
Grant Agreement Number 247228.
• Science Foundation Ireland
Grant No. SFI/08/CE/I1380 (Líon-2)
• COST Action ICO801 on Agreement Technologies
Short-term scientific mission (STSM 1868)
• Upcoming: SFI Travel Supplement
31
32. Thanks for your attention!
• Questions?
• Contacts:
– Jodi Schneider jodi.schneider@deri.org
– Adam Wyner azwyner@liverpool.ac.uk
– Katie Atkinson katie@liverpool.ac.uk
– Trevor Bench-Capon tbc@liverpool.ac.uk
April 20, 2012 London Argumentation Forum
Notes de l'éditeur
15-15:30Identifying Arguments for Evaluation using an Argument Explorer Jodi Schneider+, Adam Wyner*, Katie Atkinson*, and Trevor Bench-Capon* * University of Liverpool +DERI, NUI Galway Argumentation is key to understanding and evaluating many texts, such as opinionated reviews, scientific articles, and persuasive blog posts. However, first the arguments in the texts must be identified, and so far, identifying and diagramming arguments with current tools (e.g. Araucaria or Rationale) has required substantial work from human analysts. With automatic text analysis, we can save time, make argument identification more objective, and speed the work of human analysts by highlighting potential argumentative sections of a text according to indicative generic argument terms (e.g. 'suppose' or 'therefore') or specific terms found in argumentation schemes (e.g. 'expert' or 'fairness'). In addition, domain terminology may be used to localise topical argument elements, that is, what the argument is about. From a corpus of Amazon camera reviews, we are developing a tool -- an Argument Explorer -- using the General Architecture for Text Engineering system, which supports a user in identifying and extracting the arguments about products from product reviews. By helping analysts to more quickly parse arguments out of texts, we would thus enable more arguments to be extracted, abstracted, and passed downstream to argument evaluation tools such as ASPARTIX or Carneades for evaluation.
See also S. Heras, K. Atkinson, V. J. Botti, F. Grasso, V. Julia ́n, and P. McBurney. How argumentation can enhance dialogues in social networks. In P. Baroni, F. Cerutti, M. Giacomin, and G. R. Simari, editors, Proceedings of COMMA ’10, pages 267–274. IOS Press, 2010.
See also S. Heras, K. Atkinson, V. J. Botti, F. Grasso, V. Julia ́n, and P. McBurney. How argumentation can enhance dialogues in social networks. In P. Baroni, F. Cerutti, M. Giacomin, and G. R. Simari, editors, Proceedings of COMMA ’10, pages 267–274. IOS Press, 2010.
See also S. Heras, K. Atkinson, V. J. Botti, F. Grasso, V. Julia ́n, and P. McBurney. How argumentation can enhance dialogues in social networks. In P. Baroni, F. Cerutti, M. Giacomin, and G. R. Simari, editors, Proceedings of COMMA ’10, pages 267–274. IOS Press, 2010.
Colors represent annotations in the text. We can then search for a large body of text
Screenshot from GATE, in which we have built components of a toolPurple: conclusionOrange: premiseLots of ambiguity – different meanings of the words*DOES* draw attention to relevant places. Can turn on & off particular things that we’re looking for. Helps with the search problem.
binary values (such as has a flash), properties with ranges (such as the number of megapixels, scope of the zoom, or lens size), and multi-slotted properties (e.g. the warranty).
Gets at values
Drawn from vast lists of terminology, given sentiment valence: positive vs. negative +5 to 0 to -5Can look for various levels or homogenize – this is homogenized
Lots more issues to consider.We do one piece at a time. Small parts start to combine into more and more elements.So far: division of topics
We have an argument for buying the camera, an argument for not buying the camera. They rebut each other.We have attacks on the premises for “don’t buy the camera”. The argument for not buying the camera is defeated; the argument for buying the camera stands. So you should buy the camera.