PyDial3: the Python Statistical Dialog System

PyDial Home Page

• Home Page

Tutorials

• To get familiar with usage of PyDial, you can use the following tutorials:

• Introduction to PyDial

• Policy module

• Spoken Language Understanding

• Natural Language Generation

• How to add your own module

• How to add a new domain

• Deep Reinforcement Learning Policies

System Documentation

• Agent - the Complete Dialogue System
  • Agent.py - wrapper for all components required in a dialogue system
• Semantic Belief Tracking
  • ModularSemanticBeliefTraker.py - separate modelling of semantic decoding and belief tracking
• Semantic Parsing
  • SemI.py - Semantic input parser
  • RuleSemIMethods.py - Semantic input parsers based on rules
  • RegexSemI.py - Regular expressions SemI parser base class
• Belief Tracking
  • BeliefTrackingManager.py - wrapper for belief tracking across domains
  • BeliefTracker.py - Belief Tracker
  • baseline.py - Baseline Belief Tracker from DST
  • BeliefTrackingUtils.py - Belief Tracking Utility Methods
• Policy
  • Policy.py - abstract class for all policies
  • PolicyManager.py - container for all policies
  • PolicyCommittee.py - implementation of the Bayesian committee machine for dialogue management
  • HDCPolicy.py - Handcrafted dialogue manager
  • GPPolicy.py - Gaussian Process policy
  • GPLib.py - Gaussian Process SARSA algorithm
  • HDCTopicManager.py - policy for the front end topic manager
  • WikipediaTools.py - basic tools to access wikipedia
  • SummaryAction.py - Mapping between summary and master actions
  • SummaryUtils.py - summarises dialog events for mapping from master to summary belief
  • PolicyUtils.py - Utility Methods for Policies
  • BCM_Tools.py - Script for creating slot abstraction mapping files
• DeepRL Policies
  • A2CPolicy.py - Advantage Actor-Critic policy
  • ACERPolicy.py - Sample Efficient Actor Critic with Experience Replay
  • DQNPolicy.py - deep Q network policy
  • ENACPolicy.py - Episodic Natural Actor-Critic policy
  • TRACERPolicy.py - Trust region advantage Actor-Critic policy with experience replay
  • FeudalGainPolicy.py - Information Gain for FeudalRL policies
• FeudalRL Policies
• Natural Language Generation
  • SemOManager.py - Semantic Output
  • RuleSemOMethods.py - Classes for all Rule based Generators
  • RNNSemOMethods.py - Interface for all RNN based Generators
• Evaluation
  • EvaluationManager.py - module for determining the reward
• Topic Tracking
  • TopicTracking.py - Topic tracking interface
  • RuleTopicTrackers.py - Rule based topic trackers

---

• Conversational Entity Dialogue Model

---

• Dialogue Systems
  • DialogueServer.py - Audio and VoIP interface to Spoken SDS
  • Simulate.py - semantic level user simulator system.
  • Texthub.py - text level dialog system.
• Simulation
  • SimulatedUsersManager.py - combines simulated components into functional simulator
  • ConfidenceScorer.py -
  • ConfusionModel.py - handcrafted SemI error creator
  • ErrorModel.py - error simulation
  • NBestGenerator.py -
  • UserModel.py - goal, agenda inventor for sim user
  • UMHdcSim.py - Handcrafted simulated user behaviour

---

• Ontology
  • DataBase.py - defines the data base interface
  • FlatOntologyManager.py - Domain class and Multidomain API
  • Ontology.py - Provides system wide access to ontology
  • OntologyUtils.py - paths and rules for ontology
• Functional Tests
  • Running PyDial Test Programs
  • List of Current Tests
• Utility Modules
  • ContextLogger.py - wrapper for Python logging API
  • dact.py - dialogue act specification
  • DiaAct.py - dialogue act specification that extends dact.py
  • DialogueState.py - dialogue state object specification
  • dummyDialogueServerClient.py - script for talking to the dialogue server via cmd line, i.e. without using any voice client.
  • Scanner.py - string scanning utility
  • Settings.py - global variables: config, random num generator

---

• Acknowledgements

Indices and tables

• Index

• Module Index

• Search Page