Calo Project Home @ CSLI - Stanford

This is the public CALO home page for the CSLI Semlab team.

We are researching robust multimodal natural language and discourse understanding for use in monitoring, recording, and summarizing multi-party meetings. Research foci include automatic topic segmentation and extraction, decision detection, multimodal fusion, ontological discourse modelling, robust semantic parsing, and dialogue act detection. Funded by DARPA through SRI.

Publications

Publications related to the CALO project are listed under CaloPublications. For a full list of our lab's publications, please see the general publications page.

Software

General Documentation

• Getting Started - General instructions for downloading, installing, developing, linking, and running our software
• 3rd-Party Software - Installation instructions for third-party supporting software
• Y4 Change Log - A log of software changes for Y4

End-User Documentation

• Meeting Browser - A mozilla-embedded meeting browser UI and feedback client
• NOMOS - A playback, visualization, and annotation tool
  • Additional instructions for CALO users

Integration "How To"...

• GUI Usage HOWTO - Instructions on using our software via the GUI
• API Usage HOWTO - Instructions on using our software via the APIs
• Y4 IPTO Release Deployment - Coming soon...

Programming Documentation

• As a preface, see the architecture diagram
• OAA and The Test - Instructions for what to run and when for answering PQs

• API Reference
  • Javadocs - Documentation for our Java software
  • ServiceManager - The main Java plugin point for meeting server applications to obtain connections to the following services:
  • RobustParserService - Provides large-vocabulary semantic parsing for speech and text
  • TopicSegmenterService - Provides topic exctraction and segmentation for meetings
  • OnlineTopicSegmenterService - Provides online topic exctraction and segmentation for meetings
  • ActionItemIdentifierService - Detects action items in meetings
  • FeedbackService - An interface for providing user-feedback to detection algorithms
  • AddresseeDetectorService - An interface for getting information about utterance addressee

Data and Annotations

• NomosOpiAnnotationInfo - contains documentation of the ontologies (schemata) used in the available datasets
• TopicActionItemDataset - Our 2004 (Y2) topic and action-item annotations of the ICSI and ISL meeting corpora

Previous Years' Reference Documentation

Y3 Info

• CaloY3SystemSummary: An overview of our Y3 software and deliverables
• CaloY3SystemDiscussion: A working document discussing our Y3 software architecture
• CaloY3ArchitectureDiagram: The proposed architecture in diagram form
• Demo Instructions - instructions on how to install and run the Y3 demo
• Y3 Change Log - A log of software changes a we approach the test
• Running time for topics - Informations on topic-related queries running time

Y3 Test

• Individual PQ Analysis - The 16-question analysis of the PQs for which we are Developer POC
  • CaloY3ActionItemTestQuestions
  • CaloY3TopicTestQuestions

Y2 Modules

• MOKB: The Meeting Ontology Knowledge Base.
• CaloAutostarter
• CaloDialogueClient
• CaloDialogueServer
• CaloGeminiNlAgent
• CaloMeetingAnalyzer
• CaloMeetingReviewer
• CaloMeetingSummarizer

Y2 Ontology Documentation

The following is the original CALO-oriented documentation of the Y2 multimodal discourse ontology. Also see our MLMI '05 paper for a cleaner high-level discussion.

• MokbOntologyY2Specification
• MokbOntologyUsageExamples
• MokbOntologyDiagram
• MokbOntologyIssues
• MokbOntologyOwlFiles
• MokbOntologyNotes
• MokbOntologyY2Notes

Y2 Test Documentation

• IetQuestionsAnalysisY3: Y3 analysis of quesions
• NewIetQuestionsY3: Analysis of the CSLI-relevant IET questions
• IetQuestionsAnalysisY2: Y2 analysis of quesions