Topic Trees And Chatbot Conversational Context

Abstract

This article discusses the relation between topic-specific word trees (“ontologies”) and the retrieval of information from a chatbot‟s recent conversation – the conversational context. I assume the conversational context includes the sequence of recently visited nodes of the topic tree. So we caninterpret context retrieval as a problem of selection from this sequence, in reverse order. We will see that a surprisingly small set of selection methods support a diversity of indeterminate words and phrasesthat may need to be handled by a chatbot: like “it”, “them”, “what‟s the difference?”, “the third”, etc.An essential part of the solution is the way context is filteredto contain only the kinds of items we are expecting to find in the current input but that are missing. What is expected but missing drives context retrieval. In practice when an input sentence is encountered that is not understood and contains an indeterminate word, a chatbot may be allowed a second pass after retrieving and inserting recent context into the input.Theseideas are implemented in Python at: https://github.com/peterwaksman/Narwhal

Authors and Affiliations

Peter Waksman

Keywords

.

Related Articles

Modeling and control of PMSG based variable speed wind turbine using Proportional Resonant controller

This paper presents a controlling scheme of a permanent magnet synchronous generator based variable speed wind turbine connected to the utility grid, and a full-scale back-to-back voltage source converter to regulate the...

Cost Estimation Measurement in Software Development Based on Enhanced Layer Model

Cost estimation in the early phase of the software development helps the developers to find out that how many software requirements would be implemented during the different releases of the software. Cost serves as an im...

IMAGE COMPRESSION FOR DIFFERENT BLOCK PIXELALGORITHM FOR RANDOM IMAGE

This paper aims to proposed multi-level block truncation code (BTC) based image compression of continuous tone still image to achieve low bit rate and high quality. The algorithm has been proposed by combining bit map an...

Flow Investigation inside A Curved Square Duct

This paper presents the results of an experimental work with measurement of wall static pressure of 90°C shaped Curved duct. The test duct is made up of transparent perspex sheets to facilitate the flow visualization stu...

A Review on Software Mining: Current Trends and Methodologies

With the evolution of Software Mining, it has enabled to play a crucial role in the day to day activities .By empowering the software with the data mining methods it aids all the software developers and the managers at a...

Download PDF file
  • EP ID EP395030
  • DOI 10.9790/9622-0808015662.
  • Views 146
  • Downloads 0

How To Cite

Peter Waksman (2018). Topic Trees And Chatbot Conversational Context. International Journal of engineering Research and Applications, 8(8), 56-62. https://europub.co.uk/articles/-A-395030