SlideShare une entreprise Scribd logo
1  sur  13
Télécharger pour lire hors ligne
proScript:
Partially Ordered Scripts Generation
Keisuke Sakaguchi, Chandra Bhagavatula, Ronan Le Bras,

Niket Tandon, Peter Clark, Yejin Choi

https://proscript.allenai.org/
What is Script? Why is it important?
“a script is a stereotyped sequence of actions that defines
a well-known situation and has associated with it” 
Roger Schank and Robert Abelson (1977)
2
What is Script? Scenario:
Travel to Hawaii
What is Script? Why is it important?
Images from: https://images.app.goo.gl/B24EbZbZ8Xzw7STUA https://images.app.goo.gl/iWRLkmm3TTWGD4jJ7 https://images.app.goo.gl/WDjZtLHW3Gw36eS9A
“a script is a stereotyped sequence of actions that defines
a well-known situation and has associated with it” 
Roger Schank and Robert Abelson (1977)
3
✦Script is an essential part commonsense knowledge.
✦Script helps to represent and understand causal structure of events.
✦Script allows inference about implicit cause and effect relationship.
What is Script?
Why is it important?
Scenario:
Travel to Hawaii
Research Problem: trade-off between quality and scale 4
Quality Scalability
Induce from texts
(e.g., Chambers and Jurafsky, 2008)
- +
Event sequence alignment
(e.g., Regneri et al., 2010)
+ -
proScript + +
Our contributions 5
We collect 6.4k partially ordered scripts, proScript,
which is substantially larger than prior datasets.
With proScript, we introduced two complementary tasks
and models. (edge prediction and script generation)
We show the first time that pre-trained neural LM can be
adapted to generate partial-order scripts.
1. proScript: Crowdsourced 6.4k partial-order scripts 6
Suppose a scenario where someone wants to “travel to Hawaii”.
Q1: Describe 5 to 7 essential steps and each time duration. (Note: the order does not matter.)
decide schedule 1 hour
book a flight
go to airport
Q2. Create a flowchart of the steps
(possibly in partial order, where temporal
ordering is required only when it is necessary.)
30 minutes
1 hour
Collect “scenarios” (e.g., travel to Hawaii, bake a cake) from existing corpora and datasets.
1. DeScript (Wanzare et al., 2016) 2. VirtualHome (Puig et al., 2018) 3. ROCStories (Mostafazadeh et al., 2016)
2. Two complementary tasks and models
1. proScript Edge Prediction
7
find the cake recipe
gather the ingredients
turn on the oven
mix the ingredients
put the cake batter in the oven
bake for the right amount of time
take the cake out of the oven
Scenario: bake a cake
Given: Scenario and randomly shuffled events
2. proScript Generation
2. Two complementary tasks and models
1. proScript Edge Prediction
8
2. proScript Generation
find the cake recipe
gather the ingredients
turn on the oven
mix the ingredients
put the cake batter in the oven
bake for the right amount of time
take the cake out of the oven
Scenario: bake a cake
Given: Scenario and randomly shuffled events Given: Scenario and the number of events (to generate)
Scenario: bake a cake
Number of events: 7
2. Two complementary tasks and models
1. proScript Edge Prediction
9
2. proScript Generation
find the cake recipe
gather the ingredients
turn on the oven
mix the ingredients
put the cake batter in the oven
bake for the right amount of time
take the cake out of the oven
Scenario: bake a cake
Given: Scenario and randomly shuffled events Given: Scenario and the number of events (to generate)
Scenario: bake a cake
Number of events: 7
DAGs are flattened by DOT language
3. Generate partial-order Scripts with neural LM 10
Scenario: play the organ
Scenario: drink a glass of milk
walk to the kitchen
open the refrigerator
remove milk from refrigerator
close the refrigerator
pour milk into pot to warm a bit
pour milk into glass to drink
raise glass to lips
find sheet music to play
sit down at the organ bench set up the sheet
warm up on the organ
play the music on the organ
Evaluation: Graph Edit Distance (lower GED, the better) 11
random baseline
proScript generator
Human
0 3 6 9 12
2.78
4.73
11.3
← smaller the better
Human (2.7) < proScript generator (4.7) << Random (11.3)
proScript generator
Check out the paper for more details! 12
Pairwise comparison between Human vs. Model,
Qualitative analysis of graph edits,
Results on edge prediction task,
and a lot of other results and analysis…
Summary 13
We collect 6.4k partially ordered scripts, proScript,
which is substantially larger than prior datasets.
With proScript, we introduced two complementary tasks
and models. (edge prediction and script generation)
We show the first time that pre-trained neural LM can be
adapted to generate partial-order scripts.
Data is available:
https://proscript.allenai.org/

Contenu connexe

Similaire à EMNLP 2021 proScript (summary slides)

Escape from mediocrity
Escape from mediocrityEscape from mediocrity
Escape from mediocrityDavid Peabody
 
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...Zhenzhong Xu
 
Can’t Find Your 404s?
Can’t Find Your 404s?Can’t Find Your 404s?
Can’t Find Your 404s?Michael Nelson
 
LeakPair: Proactive Repairing of Leaks in Single Page Web Applications
LeakPair: Proactive Repairing of Leaks in Single Page Web ApplicationsLeakPair: Proactive Repairing of Leaks in Single Page Web Applications
LeakPair: Proactive Repairing of Leaks in Single Page Web ApplicationsDongsun Kim
 
Kubernetes in 30 minutes (2017/03/10)
Kubernetes in 30 minutes (2017/03/10)Kubernetes in 30 minutes (2017/03/10)
Kubernetes in 30 minutes (2017/03/10)lestrrat
 
2014 11-13-sbsm032-reproducible research
2014 11-13-sbsm032-reproducible research2014 11-13-sbsm032-reproducible research
2014 11-13-sbsm032-reproducible researchYannick Wurm
 
Diving into event-sourcing and event-driven architectures
Diving into event-sourcing and event-driven architecturesDiving into event-sourcing and event-driven architectures
Diving into event-sourcing and event-driven architecturesUnai Esteibar
 

Similaire à EMNLP 2021 proScript (summary slides) (7)

Escape from mediocrity
Escape from mediocrityEscape from mediocrity
Escape from mediocrity
 
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...
FlinkForward Asia 2019 - Evolving Keystone to an Open Collaborative Real Time...
 
Can’t Find Your 404s?
Can’t Find Your 404s?Can’t Find Your 404s?
Can’t Find Your 404s?
 
LeakPair: Proactive Repairing of Leaks in Single Page Web Applications
LeakPair: Proactive Repairing of Leaks in Single Page Web ApplicationsLeakPair: Proactive Repairing of Leaks in Single Page Web Applications
LeakPair: Proactive Repairing of Leaks in Single Page Web Applications
 
Kubernetes in 30 minutes (2017/03/10)
Kubernetes in 30 minutes (2017/03/10)Kubernetes in 30 minutes (2017/03/10)
Kubernetes in 30 minutes (2017/03/10)
 
2014 11-13-sbsm032-reproducible research
2014 11-13-sbsm032-reproducible research2014 11-13-sbsm032-reproducible research
2014 11-13-sbsm032-reproducible research
 
Diving into event-sourcing and event-driven architectures
Diving into event-sourcing and event-driven architecturesDiving into event-sourcing and event-driven architectures
Diving into event-sourcing and event-driven architectures
 

Plus de Keisuke Sakaguchi (9)

Acl18 sakaguchi
Acl18 sakaguchiAcl18 sakaguchi
Acl18 sakaguchi
 
Ijcnlp17 sakaguchi
Ijcnlp17 sakaguchiIjcnlp17 sakaguchi
Ijcnlp17 sakaguchi
 
ACL17_Sakaguchi
ACL17_SakaguchiACL17_Sakaguchi
ACL17_Sakaguchi
 
TACL16_Sakaguchi
TACL16_SakaguchiTACL16_Sakaguchi
TACL16_Sakaguchi
 
NAACL15_sakaguchi
NAACL15_sakaguchiNAACL15_sakaguchi
NAACL15_sakaguchi
 
BEA12_sakaguchi
BEA12_sakaguchiBEA12_sakaguchi
BEA12_sakaguchi
 
ACL13_sakaguchi
ACL13_sakaguchiACL13_sakaguchi
ACL13_sakaguchi
 
WMT14_sakaguchi
WMT14_sakaguchiWMT14_sakaguchi
WMT14_sakaguchi
 
COLING12_sakaguchi
COLING12_sakaguchiCOLING12_sakaguchi
COLING12_sakaguchi
 

Dernier

Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...RohitNehra6
 
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticsPulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticssakshisoni2385
 
Chromatin Structure | EUCHROMATIN | HETEROCHROMATIN
Chromatin Structure | EUCHROMATIN | HETEROCHROMATINChromatin Structure | EUCHROMATIN | HETEROCHROMATIN
Chromatin Structure | EUCHROMATIN | HETEROCHROMATINsankalpkumarsahoo174
 
Recombinant DNA technology (Immunological screening)
Recombinant DNA technology (Immunological screening)Recombinant DNA technology (Immunological screening)
Recombinant DNA technology (Immunological screening)PraveenaKalaiselvan1
 
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...Sérgio Sacani
 
DIFFERENCE IN BACK CROSS AND TEST CROSS
DIFFERENCE IN  BACK CROSS AND TEST CROSSDIFFERENCE IN  BACK CROSS AND TEST CROSS
DIFFERENCE IN BACK CROSS AND TEST CROSSLeenakshiTyagi
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSarthak Sekhar Mondal
 
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...ssifa0344
 
Chemistry 4th semester series (krishna).pdf
Chemistry 4th semester series (krishna).pdfChemistry 4th semester series (krishna).pdf
Chemistry 4th semester series (krishna).pdfSumit Kumar yadav
 
fundamental of entomology all in one topics of entomology
fundamental of entomology all in one topics of entomologyfundamental of entomology all in one topics of entomology
fundamental of entomology all in one topics of entomologyDrAnita Sharma
 
GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)Areesha Ahmad
 
Zoology 4th semester series (krishna).pdf
Zoology 4th semester series (krishna).pdfZoology 4th semester series (krishna).pdf
Zoology 4th semester series (krishna).pdfSumit Kumar yadav
 
Botany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfBotany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfSumit Kumar yadav
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)Areesha Ahmad
 
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPirithiRaju
 
Disentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTDisentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTSérgio Sacani
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisDiwakar Mishra
 
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsHubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsSérgio Sacani
 

Dernier (20)

Biopesticide (2).pptx .This slides helps to know the different types of biop...
Biopesticide (2).pptx  .This slides helps to know the different types of biop...Biopesticide (2).pptx  .This slides helps to know the different types of biop...
Biopesticide (2).pptx .This slides helps to know the different types of biop...
 
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceuticsPulmonary drug delivery system M.pharm -2nd sem P'ceutics
Pulmonary drug delivery system M.pharm -2nd sem P'ceutics
 
Chromatin Structure | EUCHROMATIN | HETEROCHROMATIN
Chromatin Structure | EUCHROMATIN | HETEROCHROMATINChromatin Structure | EUCHROMATIN | HETEROCHROMATIN
Chromatin Structure | EUCHROMATIN | HETEROCHROMATIN
 
Recombinant DNA technology (Immunological screening)
Recombinant DNA technology (Immunological screening)Recombinant DNA technology (Immunological screening)
Recombinant DNA technology (Immunological screening)
 
The Philosophy of Science
The Philosophy of ScienceThe Philosophy of Science
The Philosophy of Science
 
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
All-domain Anomaly Resolution Office U.S. Department of Defense (U) Case: “Eg...
 
CELL -Structural and Functional unit of life.pdf
CELL -Structural and Functional unit of life.pdfCELL -Structural and Functional unit of life.pdf
CELL -Structural and Functional unit of life.pdf
 
DIFFERENCE IN BACK CROSS AND TEST CROSS
DIFFERENCE IN  BACK CROSS AND TEST CROSSDIFFERENCE IN  BACK CROSS AND TEST CROSS
DIFFERENCE IN BACK CROSS AND TEST CROSS
 
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatidSpermiogenesis or Spermateleosis or metamorphosis of spermatid
Spermiogenesis or Spermateleosis or metamorphosis of spermatid
 
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
TEST BANK For Radiologic Science for Technologists, 12th Edition by Stewart C...
 
Chemistry 4th semester series (krishna).pdf
Chemistry 4th semester series (krishna).pdfChemistry 4th semester series (krishna).pdf
Chemistry 4th semester series (krishna).pdf
 
fundamental of entomology all in one topics of entomology
fundamental of entomology all in one topics of entomologyfundamental of entomology all in one topics of entomology
fundamental of entomology all in one topics of entomology
 
GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)GBSN - Microbiology (Unit 1)
GBSN - Microbiology (Unit 1)
 
Zoology 4th semester series (krishna).pdf
Zoology 4th semester series (krishna).pdfZoology 4th semester series (krishna).pdf
Zoology 4th semester series (krishna).pdf
 
Botany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdfBotany 4th semester series (krishna).pdf
Botany 4th semester series (krishna).pdf
 
GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)GBSN - Microbiology (Unit 2)
GBSN - Microbiology (Unit 2)
 
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdfPests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
Pests of cotton_Borer_Pests_Binomics_Dr.UPR.pdf
 
Disentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOSTDisentangling the origin of chemical differences using GHOST
Disentangling the origin of chemical differences using GHOST
 
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral AnalysisRaman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
Raman spectroscopy.pptx M Pharm, M Sc, Advanced Spectral Analysis
 
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroidsHubble Asteroid Hunter III. Physical properties of newly found asteroids
Hubble Asteroid Hunter III. Physical properties of newly found asteroids
 

EMNLP 2021 proScript (summary slides)

  • 1. proScript: Partially Ordered Scripts Generation Keisuke Sakaguchi, Chandra Bhagavatula, Ronan Le Bras,
 Niket Tandon, Peter Clark, Yejin Choi
 https://proscript.allenai.org/
  • 2. What is Script? Why is it important? “a script is a stereotyped sequence of actions that defines a well-known situation and has associated with it”  Roger Schank and Robert Abelson (1977) 2 What is Script? Scenario: Travel to Hawaii
  • 3. What is Script? Why is it important? Images from: https://images.app.goo.gl/B24EbZbZ8Xzw7STUA https://images.app.goo.gl/iWRLkmm3TTWGD4jJ7 https://images.app.goo.gl/WDjZtLHW3Gw36eS9A “a script is a stereotyped sequence of actions that defines a well-known situation and has associated with it”  Roger Schank and Robert Abelson (1977) 3 ✦Script is an essential part commonsense knowledge. ✦Script helps to represent and understand causal structure of events. ✦Script allows inference about implicit cause and effect relationship. What is Script? Why is it important? Scenario: Travel to Hawaii
  • 4. Research Problem: trade-off between quality and scale 4 Quality Scalability Induce from texts (e.g., Chambers and Jurafsky, 2008) - + Event sequence alignment (e.g., Regneri et al., 2010) + - proScript + +
  • 5. Our contributions 5 We collect 6.4k partially ordered scripts, proScript, which is substantially larger than prior datasets. With proScript, we introduced two complementary tasks and models. (edge prediction and script generation) We show the first time that pre-trained neural LM can be adapted to generate partial-order scripts.
  • 6. 1. proScript: Crowdsourced 6.4k partial-order scripts 6 Suppose a scenario where someone wants to “travel to Hawaii”. Q1: Describe 5 to 7 essential steps and each time duration. (Note: the order does not matter.) decide schedule 1 hour book a flight go to airport Q2. Create a flowchart of the steps (possibly in partial order, where temporal ordering is required only when it is necessary.) 30 minutes 1 hour Collect “scenarios” (e.g., travel to Hawaii, bake a cake) from existing corpora and datasets. 1. DeScript (Wanzare et al., 2016) 2. VirtualHome (Puig et al., 2018) 3. ROCStories (Mostafazadeh et al., 2016)
  • 7. 2. Two complementary tasks and models 1. proScript Edge Prediction 7 find the cake recipe gather the ingredients turn on the oven mix the ingredients put the cake batter in the oven bake for the right amount of time take the cake out of the oven Scenario: bake a cake Given: Scenario and randomly shuffled events 2. proScript Generation
  • 8. 2. Two complementary tasks and models 1. proScript Edge Prediction 8 2. proScript Generation find the cake recipe gather the ingredients turn on the oven mix the ingredients put the cake batter in the oven bake for the right amount of time take the cake out of the oven Scenario: bake a cake Given: Scenario and randomly shuffled events Given: Scenario and the number of events (to generate) Scenario: bake a cake Number of events: 7
  • 9. 2. Two complementary tasks and models 1. proScript Edge Prediction 9 2. proScript Generation find the cake recipe gather the ingredients turn on the oven mix the ingredients put the cake batter in the oven bake for the right amount of time take the cake out of the oven Scenario: bake a cake Given: Scenario and randomly shuffled events Given: Scenario and the number of events (to generate) Scenario: bake a cake Number of events: 7 DAGs are flattened by DOT language
  • 10. 3. Generate partial-order Scripts with neural LM 10 Scenario: play the organ Scenario: drink a glass of milk walk to the kitchen open the refrigerator remove milk from refrigerator close the refrigerator pour milk into pot to warm a bit pour milk into glass to drink raise glass to lips find sheet music to play sit down at the organ bench set up the sheet warm up on the organ play the music on the organ
  • 11. Evaluation: Graph Edit Distance (lower GED, the better) 11 random baseline proScript generator Human 0 3 6 9 12 2.78 4.73 11.3 ← smaller the better Human (2.7) < proScript generator (4.7) << Random (11.3) proScript generator
  • 12. Check out the paper for more details! 12 Pairwise comparison between Human vs. Model, Qualitative analysis of graph edits, Results on edge prediction task, and a lot of other results and analysis…
  • 13. Summary 13 We collect 6.4k partially ordered scripts, proScript, which is substantially larger than prior datasets. With proScript, we introduced two complementary tasks and models. (edge prediction and script generation) We show the first time that pre-trained neural LM can be adapted to generate partial-order scripts. Data is available: https://proscript.allenai.org/