Chatbot

from Wikipedia, the free encyclopedia

A chatterbot , chatbot or bot for short is a text-based dialogue system that allows chatting with a technical system. It has an area for text input and output, which can be used to communicate with the system in natural language. Chatbots can, but do not have to, be used in conjunction with an avatar . Technically, bots are more closely related to a full-text search engine than to artificial or even natural intelligence. With the increasing computer performance, however, chatbot systems can access ever more extensive databases faster and faster and therefore also offer intelligent dialogues for the user. Such systems are also known as virtual personal assistants .

There are also chatbots that do not even try to act like a human chatter (hence no chat ter bots), but rather only react to special commands, similar to IRC services . They can serve as an interface to services outside the chat, or offer functions only within your chat room, e.g. B. Greet new chatters with the joke of the day.

Today, chatbots are mostly accessed by digital assistants such as Google Assistant and Amazon Alexa , via messenger apps such as Facebook Messenger or WhatsApp, or via organizational tools and websites.

history

The history of chatbots goes back to the 1960s. Eliza , a first demonstration by a virtual psychotherapist programmed by Joseph Weizenbaum between 1964 and 1966, is considered the first chatbot .

In the decades that followed, numerous developers used Weizenbaum's model to further develop human-like interactions with chatbots.

A common goal of many who work on chatbots is to pass the Turing test .

From 2001 to 2015, the Chatterbox Challenge was held , an international competition that chose the chatbot of the year .

functionality

Most chatbots use a ready-made database, the so-called knowledge database, with answers and recognition patterns. The program first breaks down the entered question into individual parts and processes them according to given rules. Spellings can be harmonized (upper and lower case, umlauts, etc.), punctuation marks interpreted and typos can be compensated (preprocessing). In the second step, the question is actually identified. This is usually solved using recognition patterns; some chatbots also allow the nesting of various pattern recognitions using so-called macros. If an answer that matches the question is recognized, it can still be adjusted (for example, script-controlled calculated data can be inserted - “In Ulm it is 37 ° C today”). This process is called post processing. The resulting answer is then output. Modern commercial chatbot programs also allow direct access to all processing via built-in scripting languages ​​and programming interfaces .

Establishing a chatbot

This article or the following section is not adequately provided with supporting documents ( e.g. individual evidence ). Information without sufficient evidence could be removed soon. Please help Wikipedia by researching the information and including good evidence.

The challenge in programming a chatbot lies in the sensible combination of the detections. Precise detections for special questions are supplemented by global detections that only relate to one word and can serve as fallback (the bot roughly recognizes the topic, but not the exact question). Some chatbot programs support development through prioritization ranks that can be assigned to individual answers. For the programming of a chatbot, development environments are usually used that allow questions to be categorized, answers to be prioritized and recognitions to be managed. Some also allow the creation of a conversation context that is based on recognitions and possible follow-up recognitions (“Would you like to find out more about this?”). Once the knowledge base has been established, the bot is optimized in as many training sessions as possible with users in the target group. Incorrect recognitions, recognition gaps and missing answers can be identified in this way. Most of the time, the development environment offers analysis tools to efficiently evaluate the conversation protocols. In this way, a good chatbot achieves an average recognition rate of more than 70% of the questions. It is accepted by most users as an entertaining counterpart.

Multimedia chatbots

Originally purely text-based, chatbots have evolved through ever-increasing speech recognition and speech synthesis and, in addition to pure text dialogues, also offer fully spoken dialogues or a mix of both. Other media can also be used, for example images and videos. With the heavy use of mobile devices ( smartphones , wearables ), this possibility of using chatbots will continue to increase (status: Nov. 2016). As improvements continue, chatbots are not limited to just a few restricted subject areas (weather forecast, news, etc.), but enable extended dialogues and services for the user. These develop into intelligent personal assistants .

Web links

literature

  • Alexander Braun: Chatbots in Customer Communication, SpringerVS, Wiesbaden 2013, ISBN 978-3-642-62411-7 .
  • Markus Kaiser , Aline-Florence Buttkereit, Johanna Hagenauer: Chatbots. Automated communication in journalism and in public relations, SpringerVS, Wiesbaden 2019, ISBN 978-3-658-25493-3

Individual evidence

  1. ↑ Darren Orf: Google Assistant Is a Mega AI Bot That Wants To Be Absoutely Everywhere . In: Gizmodo . ( gizmodo.com [accessed June 19, 2018]).
  2. ↑ The 8 best chatbots of 2016 . In: VentureBeat . December 21, 2016 ( venturebeat.com [accessed June 19, 2018]).
  3. ↑ a b Stephanie: The history and development of chatbots. In: Onlim.de. June 11, 2020, accessed June 11, 2020 .
  4. ^ Archive access to the Chatterbox Challenge award winners. Domain was abandoned in 2016. Archived from the original on August 3, 2015 ; accessed on November 16, 2016 .
  5. ↑ Botsociety: design, preview and prototype your next chatbot or voice assistant. Accessed June 19, 2018 .
  6. ↑ Botmock - Free chatbot conversation prototyping. Retrieved June 19, 2018 (American English).
  7. ^ Watson Natural Language Understanding. November 28, 2016, accessed June 19, 2018 .
  8. ↑ What are the most common words your bot receives or sends? In: Dashbot . October 31, 2017 ( dashbot.com [accessed June 19, 2018]).  ( Page no longer available , search in web archives )  Info: The link was automatically marked as defective. Please check the link according to the instructions and then remove this notice.@1@ 2Template: Dead Link / blog.dashbot.com  
  9. ↑ Michael Yuan: Building Intelligent, Cross-platform, Messaging Bots . Addison-Wesley, 2015, ISBN 978-0-13-465061-6 ( google.co.in [accessed June 19, 2018]).
  10. ↑ Thomas Kuhn: Voice assistants change our lives. WirtschaftsWoche, July 28, 2015, accessed on November 16, 2016 .