Are You Human – Adapting and Evaluating theBot Interaction Patterns Towards the UserExpectation

Abstract

A rapidly growing number of domains employ chatbots to strengthen communication with users. Systems have evolved in such a way that more and more users struggle to detect whether they talk to another human or a machine. The distinction between human and machine, on the other hand, guides users' expectations. Assuming they are talking to a fellow human, users will inevitably be disappointed as the machine fails to understand their information needs. We study a large-scale chatbot system's dialogs. The analysis detects dissatisfaction caused by incorrect expectations concerning the chatbot's abilities. We propose to tackle this problem by better clarifying the system's capabilities.

@inproceedings{BotInteractionsPatterns2020,
author = {Andreas Lommatzsch and Danuta Ploch and Benjamin Kille and Sahin Albayrak},
title = {Are You Human - Adapting and Evaluating the Bot Interaction Patterns Towards the User Expectation},
conference = {CONVERSATIONS2020 - 4th International Workshop on Chatbot Research, Amsterdam, the Netherlands},
year = {2020},
numpages = {6},
date = {November 23-24, 2020},
note = {https://conversations2020.files.wordpress.com/2020/11/conversations_2020_paper12_lommatzsch.pdf};
}
Autoren:
Kategorie:
Poster Paper
Jahr:
2020
Ort:
CONVERSATIONS 2020, 4th International Workshop on Chatbot Research, Amsterdam, the Netherlands, November 23-24, 2020
Link: