Title | Slides CAT |
---|---|
Course | CAT |
Institution | Universitat Pompeu Fabra |
Pages | 15 |
File Size | 1.5 MB |
File Type | |
Total Downloads | 57 |
Total Views | 161 |
Ejemplo de diapositivas para la presentación del articulo de CAT...
MACHINE LEARNING RAUL BARRERA GARCIA
Communication in Technical English 2nd December 2015
OVERVIEW • Introduction to Machine Learning • The concept being real • The inflexion points of Machine Learning
INTRODUCTION TO MACHINE LEARNING
INTRODUCTION TO MACHINE LEARNING (I) • Mechanic calculator appeared on 17th century [1]. • First computers introduce human intelligence inside technology.
• “Computing
machinery and intelligence” article of Alan Turing [2].
• Appearance of Artificial Intelligence (AI). • Technologies are important in our society.
Alan Turing [3]
[1] Wikimedia foundation Inc., 2015-11-03, Mechanical calculator, https://en.wikipedia.org/w/index.php?title=Mechanical_calculator&action=history, Wikipedia, accessed 2015-11-04. [2] Turing A.M., 1950, Computing machinery and intelligence, Mind, pp. 433-460. [3] http://a2.files.biography.com/image/upload/c_fill,cs_srgb,dpr_1.0,g_face,h_300,q_80,w_300/MTE5NDg0MDU1MTUzMTE2Njg3.jpg, accessed 2015-12-01.
INTRODUCTION TO MACHINE LEARNING (II) “Machine Learning (ML) is the ability of computers to learn without being programmed”- Arthur Samuel [4-5].
Arthur Samuel [6]
[4] Simon P., 2013, Too Big to Ignore: The Business Case for Big Data, Wiley. pp. 89. [5] Wikimedia foundation Inc., 2015-11-04, Machine learning, https://en.wikipedia.org/wiki/Machine_learning Wikipedia, accessed 2015-11-04. [6] University of Stanford, http://www-cs.stanford.edu/sites/default/files/Arthur%20Samuel%20picture.jpg, accessed 2015-12-01.
THE CONCEPT BEING REAL
THE CONCEPT BEING REAL (I) • Machine Learning (ML) is a probabilistic learning [7]. • Humans learn in different ways and machines are not an exception. • Relevant Types of ML: • • • • • •
Deductive ML: takes a new knowledge derived from known knowledge [8-9]. Inductive ML: apply the inductive procedure [8-9]. Analytic ML: get new knowledge analyzing information [8-9]. Analogic ML: get new knowledge comparing information [8-9].
Connectionist ML: it is built like adaptable neural network [8-9]. Supervised, not supervised, reinforcement learning, etc. [8-9].
[7] Caparrini F.S., 2015-07-16, Introducción al Aprendizaje Automático, http://www.cs.us.es/~fsancho/?e=75, Fernando Sancho Caparrini, in Spanish, accessed 2015-11-04. [8] Moreno, A., Aprendizaje automático. Edicions UPC, 1994, http://upcommons.upc.edu/bitstream/handle/2099.3/36157/9788483019962.pdf?sequence=1&isAllowed=y , in Spanish, accessed 2015-11-04. [9] Béjar J., 2012, Introduction to Machine Learning, http://www.cs.upc.edu/~bejar/apren/docum/trans/00-introAprendizaje-eng.pdf, accessed 2015-11-05.
THE CONCEPT BEING REAL (II) • Machine Learning (ML) is real and common in advanced society. • ML has many applications: • Economy and information industry: big data, data mining, make prediction [5] [10-11]. • Biology and medical industry: make diagnosis, pattern recognition or health monitoring [5] [10-11].
• Image and audio industry: processing image and audios [5] [10-11]. • Security industry [5] [10-11]. • … And more.
[5] Wikimedia foundation Inc., 2015-11-04, Machine learning, https://en.wikipedia.org/wiki/Machine_learning Wikipedia, accessed 2015-11-04. [10] Aram, 2012-03-7, Áreas de aplicación de machine learning, http://tesis-aram-dcc.blogspot.com.es/2012/03/areas-de -aplicacion-de-machine-learning.html, in Spanish, accessed 2015-1105. [11] Yeomans J., 2015-07-7, What Every Manager Should Know About Machine Learning https://hbr.org/2015/07/what-every-manager-should-know-about-machine-learning, accessed 201511-05.
THE CONCEPT BEING REAL (III) • Some examples of Machine Learning: • University of Gothenburg developed a robot that imitates the learning of a children [12].
• Google tries to launch quantum computers [13]. • IBM developed a cognitive system call Watson [14].
Watson cognitive system [15].
[12] University of Gothenburg, 2013-9-23, Artificial intelligence that imitates children’s learning, ScienceDaily, http://www.sciencedaily.com/releases/2014/09/140923085937.htm, accessed 2015-11-05. [13] Neven H., 2013-05-16, Launching the Quantum Artificial Intelligence Lab, googleresearch.blogspot.com.es/2013/05/launching-quantum-artificial.html, accessed 2015-11-05. [14] IBM Watson, 2015, http://www.ibm.com/smarterplanet/us/en/ibmwatson/ , accessed 2015-11-05. [15] Dataprix, 2014-1-20, http://www.dataprix.com/files/uploads/2799image/2014/enero%202014/Noticia_IBM_Watson.jpg.pagespeed.ce.uo1VT66lBS.jpg , accessed 2015-12-1.
THE INFLEXION POINTS OF MACHINE LEARNING
THE INFLEXION POINTS OF MACHINE LEARNING • Machine Learning is a relative new concept. • It is easy to learn by humans? • If we do not know how human works how we pretend to create an artificial human?
• Many ways to learn, but one of the best ways to learn is from mistakes. • How do you expect that it learn correctly if are not educated before. • We can lost humanity in this process of automatize our life.
SUMMARY • Machine learning (ML) is a relative know concept in history. • ML has a big impact in our society.
• ML has different ways to be implemented. • ML has a complex implementation. • ML has too many gaps to resolve before taking important decisions.
REFERENCES (I) • [1] Wikimedia foundation Inc., 2015-11-03, Mechanical calculator,
https://en.wikipedia.org/w/index.php?title=Mechanical_calculator&action=history, Wikipedia, accessed 2015-11-04.
• [2] Turing A.M., 1950, Computing machinery and intelligence, Mind, pp. 433-460. • [3] http://a2.files.biography.com/image/upload/c_fill,cs_srgb,dpr_1.0,g_face,h_300,q_80,w_300/MTE5NDg0MDU1MTUzMTE2Njg3.jpg, accessed 2015-12-01.
• [4] Simon P., 2013, Too Big to Ignore: The Business Case for Big Data, Wiley. pp. 89. • [5] Wikimedia foundation Inc., 2015-11-04, Machine learning, https://en.wikipedia.org/wiki/Machine_learning Wikipedia, accessed 201511-04.
• [6] University of Stanford, http://www-cs.stanford.edu/sites/default/files/Arthur%20Samuel%20picture.jpg, accessed 2015-12-01. • [7] Caparrini F.S., 2015-07-16, Introducción al Aprendizaje Automático, http://www.cs.us.es/~fsancho/?e=75, Fernando Sancho Caparrini, in Spanish, accessed 2015-11-04.
• [8] Moreno, A., Aprendizaje automático. Edicions UPC, 1994,
http://upcommons.upc.edu/bitstream/handle/2099.3/36157/9788483019962.pdf?sequence=1&isAllowed=y, in Spanish, accessed 201511-04.
• [9] Béjar J., 2012, Introduction to Machine Learning, http://www.cs.upc.edu/~bejar/apren/docum/trans/00-introAprendizaje-eng.pdf, accessed 2015-11-05.
REFERENCES (II) • [10] Aram, 2012-03-7, Áreas de aplicación de machine learning http://tesis-aram-
dcc.blogspot.com.es/2012/03/areas-de-aplicacion-de-machine-learning.html, in Spanish, accessed 201511-05.
• [11] Yeomans J., 2015-07-7, What Every Manager Should Know About Machine Learning
https://hbr.org/2015/07/what-every-manager-should-know-about-machine-learning, accessed 2015-1105.
• [12] University of Gothenburg, 2013-9-23, Artificial intelligence that imitates children’s learning,
ScienceDaily, http://www.sciencedaily.com/releases/2014/09/140923085937.htm, accessed 2015-1105.
• [13] Neven H., 2013-05-16, Launching the Quantum Artificial Intelligence Lab,
googleresearch.blogspot.com.es/2013/05/launching-quantum-artificial.html, accessed 2015-11-05.
• [14] IBM Watson, 2015, http://www.ibm.com/smarterplanet/us/en/ibmwatson/, accessed 2015-11-05. • [15] Dataprix, 2014-1-20,
http://www.dataprix.com/files/uploads/2799image/2014/enero%202014/Noticia_IBM_Watson.jpg.pag espeed.ce.uo1VT66lBS.jpg, accessed 2015-12-1.
THANK YOU FOR YOUR ATTENTION ANY QUESTIONS?...