![Acknowledgement
The research presented in this poster was supported by
project No.264207, ERRIC-Empowering Romanian
Research on Intelligent Information Technologies (FP7-
REGPOT-2010-1).
Acknowledgement
The research presented in this poster was supported by
project No.264207, ERRIC-Empowering Romanian
Research on Intelligent Information Technologies (FP7-
REGPOT-2010-1).
Lost Ideas identification
- Lost ideas = concepts found only in a small number of chats
- we first identified the rare words from the corpus (the ones found only
once) and afterwards checked to see if they were on-topic by evaluating the
utterances containing them (to see whether they contained on-topic
concepts).
New Ideas/Similar ideas
- New ideas = concepts that have a high frequency in a small number of
chats.
- For each utterance where we found such a concept, we extracted a text
pattern (a window of type “word1 [*] word2”) and used this pattern to detect
similar situations in the other chats.
Identifying Ideas and Reactions
- We automatically classify the chat utterances according to their content in
the following 5 different classes:
Ignored utterances – usually they contain social noise.
Ideas – utterances containing concepts that are on-topic.
Conclusions – contain patterns that are specific to conclusive
utterances.
Developments to conclusions – if the previous utterance contains some
on-topic concepts, its development should contain similar words.
Reactions – approvals, disapprovals or continuations of previously
stated ideas.
- Needed for classifying the participants according to their discourse
creativity: leaders, developers, innovators, negativists, and conclusives.
Costin-Gabriel Chiru and Traian Rebedea
“Politehnica” University of Bucharest, Department of Computer Science and
Engineering
costin.chiru@cs.pub.ro, traian.rebedea@cs.pub.ro
Purpose
- Assessing discourse creativity based on processing
a large volume of chat conversations on the same
topic.
Assumptions
- Rare concepts (that are not off-topic) as being
intentional deviations from the standard concepts
representing a sign of discourse and linguistic
creativity because they are related to the subject
matter and are expressed only by a very small
percentage of the participants to these conversations
on the same topic.
- Creativity is linked to the differences in the
participants’ discourses.
Corpus
- Senior year undergraduate students in Computer
Science studying HCI class grouped in small teams (4-
8 students/chat).
- 62 chat conversations, debating about the utility and
limitations of the existing collaborative web tools (e.g.
forums, blogs, chats, wikis, google wave, etc.) for
finding the best instrument for information sharing and
collaboration.
Creativity influenced by
- The conversation content (the new ideas and to see
what happened to them).
-The participants typology (the character of each
participant is very important for the outcome of the
conversation and the learning process).
-The gender distribution for each of the identified
typology.
Examples of Lost Ideas
- indeed you are talking about a solution that can expand can save
all the contributions that a single user has to that particular blog but
again it’s the era of speed so synchronous communication is more
important
- but you can't have a coherent conversation in real life face to face
with 5 people at the same time if everyone is talking the same thing
is true for Internet chat everybody is talking and you can't make
anything out of it
- messenger is more of a social tool not very good for cooperating
between a large number of people
- and i think chat is mostly for direct person-to-person
communication i found it rather difficult to talk in a chat conference
because it lacks organization
New Ideas
-52 different patterns: technology[*]forum, blog[*]online,
opinion[*]topic, opinion[*]information, forum[*]blog, chat[*]blog,
topic[*]chat, blog[*]info
Conclusions
- Discourse creativity is a concept that is very difficult to assess
automatically.
- We presented a method for determining creativity in a large
collection of online conversations that debate the same topic within a
CSCL scenario.
- The proposed method can be used to identify typologies of
participants starting from the elements used for assessing creativity.
References
- Rebedea, T., Dasc lu, M., Trausan-Matu, Armitt, G., and Chiru, C. Automatică
Assessment of Collaborative Chat Conversations with PolyCAFe. In:
Proceedings of ECTEL 2011, LNCS 6964, Springer, pp. 299-312 (2011).
- Renouf, A. Tracing lexical productivity and creativity in the british media: the
chavs and the chav-nots. In: Lexical Creativity, Texts and Contexts, John
Benjamins Publishing Company, Amsterdam, pp. 61-89 (2007).
- Veale, T. Creative language retrieval: A robust hybrid of information retrieval
and linguistic creativity. In: Proceedings of ACL 2011, Portland, Oregon, USA,
pp. 278-287 (2011).
-Zhu, X., Xu, Z. and Khot, T. How creative is your writing? a linguistic creativity
measure from computer science and cognitive psychology perspectives. In:
Proceedings of the Workshop on Computational Approaches to Linguistic
Creativity, 87-93, ACL (2009)
Detecting Discourse Creativity in Chat Conversations](https://image.slidesharecdn.com/detectingdiscoursecreativityinchatconversationsv4-190226144953/85/Detecting-discourse-creativity-in-chat-conversations-1-320.jpg)
Detecting discourse creativity in chat conversations
- 1. Acknowledgement The research presented in this poster was supported by project No.264207, ERRIC-Empowering Romanian Research on Intelligent Information Technologies (FP7- REGPOT-2010-1). Acknowledgement The research presented in this poster was supported by project No.264207, ERRIC-Empowering Romanian Research on Intelligent Information Technologies (FP7- REGPOT-2010-1). Lost Ideas identification - Lost ideas = concepts found only in a small number of chats - we first identified the rare words from the corpus (the ones found only once) and afterwards checked to see if they were on-topic by evaluating the utterances containing them (to see whether they contained on-topic concepts). New Ideas/Similar ideas - New ideas = concepts that have a high frequency in a small number of chats. - For each utterance where we found such a concept, we extracted a text pattern (a window of type “word1 [*] word2”) and used this pattern to detect similar situations in the other chats. Identifying Ideas and Reactions - We automatically classify the chat utterances according to their content in the following 5 different classes: Ignored utterances – usually they contain social noise. Ideas – utterances containing concepts that are on-topic. Conclusions – contain patterns that are specific to conclusive utterances. Developments to conclusions – if the previous utterance contains some on-topic concepts, its development should contain similar words. Reactions – approvals, disapprovals or continuations of previously stated ideas. - Needed for classifying the participants according to their discourse creativity: leaders, developers, innovators, negativists, and conclusives. Costin-Gabriel Chiru and Traian Rebedea “Politehnica” University of Bucharest, Department of Computer Science and Engineering costin.chiru@cs.pub.ro, traian.rebedea@cs.pub.ro Purpose - Assessing discourse creativity based on processing a large volume of chat conversations on the same topic. Assumptions - Rare concepts (that are not off-topic) as being intentional deviations from the standard concepts representing a sign of discourse and linguistic creativity because they are related to the subject matter and are expressed only by a very small percentage of the participants to these conversations on the same topic. - Creativity is linked to the differences in the participants’ discourses. Corpus - Senior year undergraduate students in Computer Science studying HCI class grouped in small teams (4- 8 students/chat). - 62 chat conversations, debating about the utility and limitations of the existing collaborative web tools (e.g. forums, blogs, chats, wikis, google wave, etc.) for finding the best instrument for information sharing and collaboration. Creativity influenced by - The conversation content (the new ideas and to see what happened to them). -The participants typology (the character of each participant is very important for the outcome of the conversation and the learning process). -The gender distribution for each of the identified typology. Examples of Lost Ideas - indeed you are talking about a solution that can expand can save all the contributions that a single user has to that particular blog but again it’s the era of speed so synchronous communication is more important - but you can't have a coherent conversation in real life face to face with 5 people at the same time if everyone is talking the same thing is true for Internet chat everybody is talking and you can't make anything out of it - messenger is more of a social tool not very good for cooperating between a large number of people - and i think chat is mostly for direct person-to-person communication i found it rather difficult to talk in a chat conference because it lacks organization New Ideas -52 different patterns: technology[*]forum, blog[*]online, opinion[*]topic, opinion[*]information, forum[*]blog, chat[*]blog, topic[*]chat, blog[*]info Conclusions - Discourse creativity is a concept that is very difficult to assess automatically. - We presented a method for determining creativity in a large collection of online conversations that debate the same topic within a CSCL scenario. - The proposed method can be used to identify typologies of participants starting from the elements used for assessing creativity. References - Rebedea, T., Dasc lu, M., Trausan-Matu, Armitt, G., and Chiru, C. Automatică Assessment of Collaborative Chat Conversations with PolyCAFe. In: Proceedings of ECTEL 2011, LNCS 6964, Springer, pp. 299-312 (2011). - Renouf, A. Tracing lexical productivity and creativity in the british media: the chavs and the chav-nots. In: Lexical Creativity, Texts and Contexts, John Benjamins Publishing Company, Amsterdam, pp. 61-89 (2007). - Veale, T. Creative language retrieval: A robust hybrid of information retrieval and linguistic creativity. In: Proceedings of ACL 2011, Portland, Oregon, USA, pp. 278-287 (2011). -Zhu, X., Xu, Z. and Khot, T. How creative is your writing? a linguistic creativity measure from computer science and cognitive psychology perspectives. In: Proceedings of the Workshop on Computational Approaches to Linguistic Creativity, 87-93, ACL (2009) Detecting Discourse Creativity in Chat Conversations