• Corpus ID: 24831380

Building a Conversational Agent Overnight with Dialogue Self-Play

  title={Building a Conversational Agent Overnight with Dialogue Self-Play},
  author={Pararth Shah and Dilek Z. Hakkani-T{\"u}r and G{\"o}khan T{\"u}r and Abhinav Rastogi and Ankur Bapna and Neha Nayak Kennard and Larry Heck},
We propose Machines Talking To Machines (M2M), a framework combining automation and crowdsourcing to rapidly bootstrap end-to-end dialogue agents for goal-oriented dialogues in arbitrary domains. [] Key Method In the first phase, a simulated user bot and a domain-agnostic system bot converse to exhaustively generate dialogue "outlines", i.e. sequences of template utterances and their semantic parses.

Figures and Tables from this paper

Bootstrapping a Neural Conversational Agent with Dialogue Self-Play, Crowdsourcing and On-Line Reinforcement Learning

This paper discusses the advantages of this approach for industry applications of conversational agents, wherein an agent can be rapidly bootstrapped to deploy in front of users and further optimized via interactive learning from actual users of the system.

Towards Scalable Multi-domain Conversational Agents: The Schema-Guided Dialogue Dataset

This work introduces the the Schema-Guided Dialogue (SGD) dataset, containing over 16k multi-domain conversations spanning 16 domains, and presents a schema-guided paradigm for task-oriented dialogue, in which predictions are made over a dynamic set of intents and slots provided as input.

MMConv: An Environment for Multimodal Conversational Search across Multiple Domains

The Multimodal Multi-domain Conversational dataset (MMConv) is introduced, a fully annotated collection of human-to-human role-playing dialogues spanning over multiple domains and tasks and adopted the state-of-the-art methods for these tasks respectively.

State-Machine-Based Dialogue Agents with Few-Shot Contextual Semantic Parsers

A methodology and toolkit for creating a rule-based multi-domain conversational agent for transactions from language annotations of the domains' database schemas and APIs and a couple of hundreds of annotated human dialogues, which achieves over 71% turn-by-turn slot accuracy on the MultiWOZ benchmark.

A framework to co-optimize task and social dialogue policies using Reinforcement Learning

This work describes a framework that allows for a RL-based agent to be able to adapt its dialogue policy depending on its user's conversational goals and relies on this result to inform the design of a social reward function that is based on an hybrid approach of supervised learning and reinforcement learning.

Dialogic: Controllable Dialogue Simulation with In-Context Learning

Experimental results on the MultiWOZ dataset demonstrate that training a model on the simulated dialogues leads to even better performance than using the same amount of human-generated dialogues under the challenging low-resource settings, with as few as 85 dialogues as a seed.

Towards a Universal NLG for Dialogue Systems and Simulators with Future Bridging

A prototype FBNLG is evaluated to show that future bridging can be a viable approach to a universal few-shot NLG for task-oriented and chit-chat dialogues.

Controllable Dialogue Simulation with In-Context Learning

A novel method for Dialogue simulation based on language model In-Context learning, dubbed as D IALOGIC, which can be used to rapidly expand a small set of dialogue data without requiring human involvement or parameter update, and is thus much more cost-efficient and time-saving than crowdsourcing.

State Machine Based Human-Bot Conversation Model and Services

The idea of leveraging Conversational State Machine to make it a core part of chatbots’ conversation engine by formulating conversations as a sequence of states is proposed, which allows chatbots to manage tangled conversation situations where most existing chatbot technologies fail.

A Few-Shot Semantic Parser for Wizard-of-Oz Dialogues with the Precise ThingTalk Representation

A new dialogue representation and a sample-efficient methodology that can predict precise dialogue states in WOZ conversations are proposed and extended the ThingTalk representation to capture all information an agent needs to respond properly.



Frames: a corpus for adding memory to goal-oriented dialogue systems

A rule-based baseline is proposed and the frame tracking task is proposed, which consists of keeping track of different semantic frames throughout each dialogue, and the task is analysed through this baseline.

A Network-based End-to-End Trainable Task-oriented Dialogue System

This work introduces a neural network-based text-in, text-out end-to-end trainable goal-oriented dialogue system along with a new way of collecting dialogue data based on a novel pipe-lined Wizard-of-Oz framework that can converse with human subjects naturally whilst helping them to accomplish tasks in a restaurant search domain.

End-to-End Optimization of Task-Oriented Dialogue Model with Deep Reinforcement Learning

A neural network based task-oriented dialogue system that can be optimized end-to-end with deep reinforcement learning (RL) and shows that deep RL based optimization leads to significant improvement on task success rate and reduction in dialogue length comparing to supervised training model.

To Plan or not to Plan? Discourse Planning in Slot-Value Informed Sequence to Sequence Models for Language Generation

This work investigates sequence-to-sequence (seq2seq) models in which slot values are included as part of the input sequence and the output surface form, and investigates whether a separate sentence planning module that decides on grouping of slot value mentions as input to the seq2seq model results in more natural sentences than a seq1seq model that aims to jointly learn the plan and the surface realization.

Sequential Dialogue Context Modeling for Spoken Language Understanding

A novel approach for modeling dialogue context in a recurrent neural network (RNN) based language understanding system that allows encoding context from the dialogue history in chronological order and results in reduced semantic frame error rates.

Scalable multi-domain dialogue state tracking

A novel framework for state tracking is introduced which is independent of the slot value set, and represent the dialogue state as a distribution over a set of values of interest (candidate set) derived from the dialogue history or knowledge, which addresses the problem of slot-scalability.

Task Completion Platform: A self-serve multi-domain goal oriented dialogue platform

A multi-domain dialogue platform that can host and execute large numbers of goal-orientated dialogue tasks, and features a task configuration language, TaskForm, that allows the definition of each individual task to be decoupled from the overarching dialogue policy used by the platform to complete those tasks.

PyDial: A Multi-domain Statistical Dialogue System Toolkit

PyDial is an opensource end-to-end statistical spoken dialogue system toolkit which provides implementations of statistical approaches for all dialogue system modules and has been extended to provide multidomain conversational functionality.

Agenda-Based User Simulation for Bootstrapping a POMDP Dialogue System

This paper investigates the problem of bootstrapping a statistical dialogue manager without access to training data and proposes a new probabilistic agenda-based method for simulating user behaviour and shows that the learned policy was highly competitive, with task completion rates above 90%.

On-line Active Reward Learning for Policy Optimisation in Spoken Dialogue Systems

An on-line learning framework whereby the dialogue policy is jointly trained alongside the reward model via active learning with a Gaussian process model is proposed.