Você está na página 1de 23

Artificial Intelligence

Wester Zela
Week One
edzela@yahoo.com
www.ssyspe.org
Resume

• Phd. Computer Science – Artificial Intelligence – Univ. Politécnica de


Madrid
• Master Software Engineering – Univ. Oxford
• Master Financial and Economic Analysis – Univ. Complutense de
Madrid
• System Engineering– Univ. Nacional Ingeniería (Perú)
• Winner of Robotrader-2014 (Automatic Trading).
• Creator of www.ssyspe.org
• 20+ experience in IT, and currently Director at HITSS Perú
Agenda

• What is intelligence/What is Artificial Intelligence?


• AI specialized branches
• Artificial Intelligence in Economy/Society
• Applications of Artificial Intelligence
• Application of AI to Computer Go
• Application of AI to Trading Systems
Course Overview
• Revisión del Silabus
• Horario:
• Sábados de 2:15pm- 6:15pm
Intelligence/Artificial Intelligence?
• Intelligence?
• One’s capacity for logic, understanding, communication, learning, memory,
problem solving,etc.
• Artificial Intelligence?
• Intelligence exhibited by machines or software.
• The study and design of intelligent agents.
• The science and engineering of making intelligent machines
https://www.youtube.com/watch?v=RsWbEA7XvOM
• Taking ideas from different fields
• Engineering, biology, psicology, communication theory, mathematics, statistics,
logic and philosophy.
Artificial Intelligence
• Robot is coming Czech word Robota: serf labor, slave work. Two Kind
of Robots:
• Science Fiction robots, look like us, think like us, behave like us: Not Exist
• Robots that does not look like us, does not behave like us. we are enveloping
the world (maps, sensors, databases, algorithms,etc): Do simple tasks ->
Success
• The central problems (or goals) of AI research include reasoning,
knowledge, planning, learning, natural language processing
(communication), perception and the ability to move and manipulate
objects.
Artificial Intelligence
• AI/ENG dept: does not matter is really AI, if It does a good job ->
Success

http://www.ssyspe.org/viewtopic.php?f=3&t=240

• AI/Cognitive Science dept:Non-Biological implementation of


Intelligence. Enourmos success if it is intelligent at all -> failing

http://www.ssyspe.org/viewtopic.php?f=11&t=395
Artificial Intelligence in Economy/Society

• AI Robots replacing humans?


• Replacing Help Desk Agents, Doctors, Teachers, Journalist, Manufacturing
workers, soldiers, etc?
• AI is bad for humanity?
• Impact in the economy, security, moral, ethic.
• Cloud Robotics
• Access to data from any robot, computation can growth, robots can talk each
other, share software (algorithms).
Artificial Intelligence - History
• 1945-1947: Turing designed ACE (Automatic Computing Engine)
• 1950: Turing Test. Simon,Newell, Cliff developed a program Logic
Theorist. Computer to Play Chess.
http://www.ssyspe.org/viewtopic.php?f=3&t=562
• 1956: First conference in AI: process symbols, new languages, etc.
• 1960: LISP, Programs ELIZA and others programs. Evolve Adaptive
Systems, Algoritmos Genéticos. Holland.
• 1960,1972: Emergence of Artificial Neural Network.
• In the early 1980s, AI research was revived by the commercial success
of expert systems, but it was very difficult found budget to research in
AI
Artificial Intelligence History
• From 1991 AI achieved greatest success with commercial applications
like logistics, data mining, medical diagnosis, etc. Reason: computer
power.
• On 11 May 1997, IBM Deep Blue became the first computer chess-
playing system to beat a world chess champion Kasparov.
https://www.youtube.com/watch?v=KF6sLCeBj0s
• In February 2011, in a Jeopardy! quiz show exhibition match, IBM's
question answering system, Watson, defeated the two greatest
Jeopardy champions.
• In March 2016, AlphaGo won 4-1 in a Go match to Go champion Lee
Sedol, becoming the first computer Go-playing system to beat a
professional human Go player without handicaps.
AI in Movies
• The Day the Earth Stood Still(1951, Robert Wise)
• 2001: A Space Odyssey(1968, Stanley Kubrick)
• Star Wars: Episode IV - A New Hope (1977, George Lucas
• The Terminator (1984, James Cameron)
• The Matrix (1999, Wachowski & Lana Wachowski)
• A.I. (2001, Steven Spielberg)
• I, Robot (2004, Alex Proyas)
• The Hitchhiker’s Guide to the Galaxy (2005, Garth Jennings)
• Her (2013, Spike Jonze)
• Avengers: Age of Ultron(2015, Joss Whedon)
• etc….
Artificial Intelligence - Branches
• AI and related topics classified in some branches:
• Expert Systems
• Natural Language Processing
• Robotics
• Computer Vision
• Machine Learning
• Genetic Algorithms
• Neural Networks
• Deep Learning
• Others
Artificial Intelligence – State of Art
• Since the beginning Computer Science and AI, games have been studied as applications fields.
Turing wrote a chess program
• Currently there are many classical games dominated by computers. i.e. checkers, Chess, tic-tac-
toe… and Go.
• Allis (94) defined the complexity of some clasical games based on Three Search. According to this
classification, a Go Board of 9x9 has the same complexity as Chess has.

• It has been applied different techniques to Games: Minimax Search, Alpha-Beta Prunning, TDL,
Montecarlo, Neuro-Evolution, Co-evolutonary algorithms
• AlphaGo of Google using Artificial Neural Network in combination with tree search.
Artificial Intelligence – State of Art
• Desde Enigma, a Big-Blue …. Watson and AlphaGo

http://www.ssyspe.org/viewtopic.php?f=11&t=311

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

https://www.youtube.com/watch?v=tCpf5wDr0UE
Some Applications of AI ….in Business
• Finance
• Banks use artificial intelligence systems to detect charges outside of norm,
invest in stocks, etc.
• Hospital and Medicine
• It is used clinical decision support systems for medical diagnosis, computer-
aided interpretation of medical images.
• Heavy Industry
• Robots are good in repetitive jobs.
• Online and telephone Customer Service
• Chatbots
• Transportation
• Cars with sensors for active security
• Toys and Games
Learning in AI
• Supervised Learning
• Agents have access to correct the test data or examples that need to be
learned. Agents can learn from the errors between its decisions and the
correct decisions.
• Unsupervised Learning
• In unsupervised learning, there’s no label or target value given for the data.
• The problem of Unsupervised Learning is try to find hidden structure in
unlabeled data.
• Reinforcement Learning
• The correct course of actions is un-known, and the agent has to learn the
good behaviour through trial and error by directly interacting with the
domain.
References
• Co-evolutionary and Reinforcement Learning Techniques in
Computer Go: Some Techniques of Machine Learning Applied to
Computer Go. Wester Zela, ISBN-10: 3659638595

• Machine Learning in Action, Peter Harrington

• Introduction to Artificial Intelligence, Wolfgang Ertel

• Reinforcement and Systemic Machine Learning for Decision Making, Parag Kulkarni
Papers of this week
• Very brief history of Artificial Intelligence, Bruce Buchanan, AI
Magazine Volume 26 Number 4, 2005.
• AlphaGo Zero: Learning from scratch
https://deepmind.com/blog/alphago-zero-learning-scratch/
• Distinguishing between Narrow AI, General AI and Super AI,
https://medium.com/@tjajal/distinguishing-between-narrow-ai-
general-ai-and-super-ai-a4bc44172e22
• The Jobs That Artificial Intelligence Will Create,
http://sloanreview.mit.edu/article/will-ai-create-as-many-jobs-as-it-
eliminates/
Final Project
• Topics:
• From repositories around the world:
• https://www.kaggle.com/datasets
• https://archive.ics.uci.edu/ml/datasets.html
• https://en.wikipedia.org/wiki/List_of_datasets_for_machine_learning_research
• https://dataverse.harvard.edu/
• http://reshare.ukdataservice.ac.uk/
• Data from other courses
• From your companies
• Work every week in Final document
• Final Presentation: 1 week before final exam
Examples of Final Projects of previous years
• Sistema para detección de caídas de ancianos
• Convolutional Neuronal Networks - Facial Emotions
• Sistema Biométrico de Reconocimiento Facial
• Sistema para reconocimiento de voz
• Chatbots
https://www.youtube.com/watch?v=qbRVndekTj0
• Algoritmo SVM para predecir el precio de acciones de Fb
http://www.ssyspe.org/viewforum.php?f=3
Final Project
• Groups?
Grupo1:
Grupo2:
Grupo3:
Grupo4:
• Topics Projects?
Project1:
Project2:
Project3:
Project4:
Ranking in ssyspe.org: 02/06
• PTOi = 2*NPi *(Vistas/Total Vistas) + 1*NPAi/Total NP
Id Friend Score Ptos
Student Post Post 02/06 02/06
1580 11.33 0.03 11.36 3.00
1573 8.30 0.04 8.34 3.00
1565 8.27 0.03 8.31 3.00
1579 2.65 0.03 2.68 2.00
1572 0.80 0.02 0.81 2.00
1571 0.53 0.04 0.57 2.00
1589 0.41 0.00 0.41 1.00
1564 0.33 0.04 0.37 1.00
1567 0.29 0.03 0.32 1.00
1566 0.15 0.02 0.18 0.00
1568 0.11 0.03 0.14 0.00
1563 0.00 0.04 0.04 0.00
1368 0.00 0.03 0.03 0.00
1574 0.00 0.03 0.03 0.00
1591 0.00 0.03 0.03 0.00
1569 0.00 0.02 0.02 0.00
1590 0.00 0.02 0.02 0.00
1570 0.00 0.01 0.01 0.00
1578 0.00 0.01 0.01 0.00
Muchas Gracias!

Você também pode gostar