diana: multilingual cross-document discourse analysis · project summary heading for:...

21
DIANA: Multilingual cross-document DIscourse ANAlysis A project proposal Maciej Ogrodniczuk Adam Przepiórkowski Linguistic Enginering Group Institute of Computer Science Polish Academy of Sciences 26 October 2012

Upload: others

Post on 01-Oct-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

DIANA:Multilingual cross-document DIscourse ANAlysis

A project proposal

Maciej Ogrodniczuk

Adam Przepiórkowski

Linguistic Enginering GroupInstitute of Computer SciencePolish Academy of Sciences

26 October 2012

Page 2: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 3: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 4: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 5: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 6: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 7: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 8: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Project summary

Heading for:

FP7-ICT-2013-10 call

Aiming at:

analysing multilingual, multimodal sources

to detect argumentation structuresin speakers’ or public opinions,

convert them into exploitable knowledge units

and provide cross-document discourse summary over time

with search capabilities and visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 2

Page 9: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Resources and results

Content:

live parliamentary corpora

live media corpora

social media (Facebook, Twitter)

Expected results:

unsupervised and semi-supervised tools for automaticdiscourse annotation for opinion detection

discussion summarizer

topic/event-based opinion search engine

visualisation methods for complex relations

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 3

Page 10: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Resources and results

Content:

live parliamentary corpora

live media corpora

social media (Facebook, Twitter)

Expected results:

unsupervised and semi-supervised tools for automaticdiscourse annotation for opinion detection

discussion summarizer

topic/event-based opinion search engine

visualisation methods for complex relations

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 3

Page 11: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Resources and results

Content:

live parliamentary corpora

live media corpora

social media (Facebook, Twitter)

Expected results:

unsupervised and semi-supervised tools for automaticdiscourse annotation for opinion detection

discussion summarizer

topic/event-based opinion search engine

visualisation methods for complex relations

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 3

Page 12: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Resources and results

Content:

live parliamentary corpora

live media corpora

social media (Facebook, Twitter)

Expected results:

unsupervised and semi-supervised tools for automaticdiscourse annotation for opinion detection

discussion summarizer

topic/event-based opinion search engine

visualisation methods for complex relations

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 3

Page 13: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Resources and results

Content:

live parliamentary corpora

live media corpora

social media (Facebook, Twitter)

Expected results:

unsupervised and semi-supervised tools for automaticdiscourse annotation for opinion detection

discussion summarizer

topic/event-based opinion search engine

visualisation methods for complex relations

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 3

Page 14: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 15: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 16: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 17: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 18: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 19: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 20: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Use case

Processing of a new manned mission to Mars in 2020 article:

topic detection → Manned mission to Mars

opinion gathering → hazards, schedule, technical issues

resolution of temporal expressions → 3 years from now

entity disambiguation → e.g. from ocean MontereyAccelerated Research System observatory

cross-document coreference resolution

discussion summarisation → in 1998 the one-way trip optionwas proposed, in 2010 Obama said that the mission willhappen by mid-2030s

visualisation

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 4

Page 21: DIANA: Multilingual cross-document DIscourse ANAlysis · Project summary Heading for: FP7-ICT-2013-10 call Aiming at: analysing multilingual, multimodal sources to detect argumentation

Call for partners

Expected to find partners proficient at:

opinion tracking

discussion analysis

Maciej Ogrodniczuk, Adam Przepiórkowski DIANA: Multilingual cross-document DIscourse ANAlysis 5