chatbots - building intelligent systems

39
www.icapps.com Chatbots: building intelligent systems Sjoera Roggeman

Upload: sjoera-roggeman

Post on 12-Apr-2017

103 views

Category:

Technology


2 download

TRANSCRIPT

Page 1: Chatbots -  building intelligent systems

www.icapps.com

Chatbots: building intelligent systemsSjoera Roggeman

Page 2: Chatbots -  building intelligent systems

WHAT ARE CHATBOTS?

Page 3: Chatbots -  building intelligent systems

“A chatbot is a service, powered by rules and sometimes artificial intelligence, that you interact with

via a chat interface”Matt Schlicht - Founder of Chatbot magazine

Page 4: Chatbots -  building intelligent systems

Poncho

Page 5: Chatbots -  building intelligent systems

Lybrate

Page 6: Chatbots -  building intelligent systems

Madison Reed

Page 7: Chatbots -  building intelligent systems

HOW DOES IT WORK?

Page 8: Chatbots -  building intelligent systems

TWO TYPES OF CHATBOTS

1. Based on rules

2. Based on Artificial

intelligence

Page 9: Chatbots -  building intelligent systems

Artificial Intelligence (AI)

“An ideal intelligent machine is a flexible rational agent that perceives its environment and takes

actions that maximize its chance of success at some goal”

Russell & Norvig, 2003

Page 10: Chatbots -  building intelligent systems

• Concept very old: Greek myths about automatons

• Beginnings of modern AI: Greek philosophers describe human thinking as a symbolic system

• Field of AI formally founded in 1956

BRIEF HISTORY OF AI

Page 11: Chatbots -  building intelligent systems

• 1997: IBM’s Deep Blue beats chess champion Garry Kasparov

BRIEF HISTORY OF AI

Page 12: Chatbots -  building intelligent systems

2011: IBM’s Watson won the quiz show Jeopardy

https://www.youtube.com/watch?v=Sp4q60BsHoY

Page 13: Chatbots -  building intelligent systems

• IBM’s Watson • Understands written and

spoken language + visuals

• Constantly learning

BRIEF HISTORY OF AI

Page 14: Chatbots -  building intelligent systems

Fields in AI

Page 15: Chatbots -  building intelligent systems

NATURAL LANGUAGE PROCESSING (NLP)

Page 16: Chatbots -  building intelligent systems
Page 17: Chatbots -  building intelligent systems

Turing Test

Page 18: Chatbots -  building intelligent systems

• Natural language understanding

• Natural language generation • Text planning • Sentence planning • Text realisation

COMPONENTS OF NLP

Page 19: Chatbots -  building intelligent systems

Steps in NLP

Lexical analysis

Syntactic analysis

Semantic analysis

Discourse integration

Pragmatic analysis

Page 20: Chatbots -  building intelligent systems

Lexical analysis

The quick brown fox jumps over the lazy dog .

article adj. adj. subst. verb prep. article adj. subst.

sentence

punct.

Page 21: Chatbots -  building intelligent systems

Syntactic analysis

The quick brown fox jumps over the lazy dog.

subjectverb adjunct

predicate

Page 22: Chatbots -  building intelligent systems

Semantic analysis

The quick brown fox jumps over the lazy dog.

Page 23: Chatbots -  building intelligent systems

Discourse integration

The quick brown fox jumps over the lazy dog.

He jumps very high.

Page 24: Chatbots -  building intelligent systems

Pragmatic analysis

The quick brown fox jumps over the lazy dog.

—> A dog is lying down, maybe sleeping (because it’s lazy). A fox takes a leap and jumps over the dog.

Page 25: Chatbots -  building intelligent systems

• “Intentions” (e.g. there’s beer in the fridge)

• Sarcasm • Irony • Ambiguity • …

—> Paul Grice’s theory of “meaning”

POSSIBLE ISSUES

• Utterer’s Meaning • Timeless Meaning

Page 26: Chatbots -  building intelligent systems

EXAMPLE

“Flying planes can be dangerous.”

Page 27: Chatbots -  building intelligent systems

NLP in practice

Page 28: Chatbots -  building intelligent systems

Example

Page 29: Chatbots -  building intelligent systems

MACHINE LEARNING (ML)

Page 30: Chatbots -  building intelligent systems

• Microsoft • Chinese market • Mines Chinese internet

for human conversations

XIAO ICE

Page 31: Chatbots -  building intelligent systems

This can also backfire!

Page 32: Chatbots -  building intelligent systems

To summarise

‘I’, ‘need’, ‘a’, ‘bunch’, ‘of’, ‘bananas’, ‘,’, ‘some’, ‘yoghurt’, ‘,’, ‘toilet’, ‘paper’, ‘,’, ‘paper’, ‘towels’, ‘1/2’, ‘lb’, ‘of’, ‘hamburger’, ‘meat’, ‘,’, ‘and’, ‘some’, ‘beer’

NLU

check for appropriate answer in database

‘By’, ‘when’, ‘?’

NLG

Page 33: Chatbots -  building intelligent systems

BUILDING A BOT

Page 34: Chatbots -  building intelligent systems

Motion.ai

Page 35: Chatbots -  building intelligent systems

Wit.ai

Page 36: Chatbots -  building intelligent systems

Google API

Page 37: Chatbots -  building intelligent systems

• One topic or several topics?

• How complex are the answers?

• What is the bot’s goal?

WHICH TOOL TO CHOOSE?

Page 38: Chatbots -  building intelligent systems

• The language is the interface

• Design with language • Cooperate with linguists,

copywriters, novelists and even comedians

TO CONCLUDE: OUR ROLE AS UX DESIGNERS?

Page 39: Chatbots -  building intelligent systems

THANK YOU!