A Historical Overview and Analysis of Artificial Intelligence - Prof. Grandal Seco, Apuntes de Periodismo

¡Descarga A Historical Overview and Analysis of Artificial Intelligence - Prof. Grandal Seco y más Apuntes en PDF de Periodismo solo en Docsity! What is artificial intelligence? Isn´t biological PRE-HISTORY The pygmalion. ARTIFICIAL INTELLIGENCE: TIMELINE • 1940s-50s: birth and first attempts to make computers copy human behavior • 1960s: General Problem Solver, Fuizziness and ELIZA • 1970s: Dark Ages: too high expectations-limited computer power- philosophers attack (Searle) • 1980s-90s: Rennaisance: e VERY BRIEF HISTORY OF AI (I) • 1943: Walter Pitts and Warren Mcculloch show how artificial neural networks can process information • 1950: Alan Turing publishes " Computing Machinery and Intelligence" • Early´s70s: invention of Prolog • Late´70s: first commercial expert systems. • ´80s: many experts systems are developed • Late´80s: growing awareness of limitations of current expert systems • Late´80s: new studies on neural networks • ´90s: Al Applications in many domains: machine learning, case based reasoning, multi-agent. • 200s: Commercial success of systems for the automated executions of legal rules (such as Ruleburst/Oracle Policy AUTOMATION) • 2010 to date: Drones, Autonomous Cars, Robots, ambient intelligence, internet of things. SUCCESSES OF AI IN... 1 Data mining: processing big quantity of data commercial purposes but also for security 2 Anti spamming technolofies 3 Medical consultation and data mining 4 Automatic translation 5 Games 6 Logistics 7 Face and movement recognition 8 Robots and virtual assistants NEGATIVE AND PESSIMISTC APPROACHES John Searle: the idea of a non-biological machine being intelligent in incoherent Hubert Dreyfus: AI is impossible Joseph Weizenbaum: the idea is obscene, anti-human and immoral Various: since AI has´t reached human level by now, it must be impossible ARTIFICIAL INTELLIGENCE (AI) • Different approaches: Biological versus phenomenal Strong and weak: Strong: create human intelligence Weak: simulate human intelligence RUSSEL AND NORVIG ON AI HUMAN OR RATIONAL AI? • Depends on wether the system is built to imitate the human or to solve a certain problem(s) in the best of ways • Human: rational and irrational at the same time • Human intelligence is much more difficult to create or simulate rather than do the same with rationality COMMON SENSE CLASSICAL AI: EXPERT SYSTEMS • Top-down approach: observe how human brain works and copy that way of functioning so as to insert it into the machine. • First approach to be used in building AI: Reasoning (if it rains and i don´t have an umbrella, i will get cold. EXPERT SYSTEMS • Experts contribute with their knowledge