*Back to the PersonalInformationStream*

The page gathers notes from the videos I have read recently (most since summer 2009). See also the ToDo section to understand how this page should evolve.

- Unsupervised Feature Learning and Deep Learning
- The Society of Mind
- 2009 AGI Summer School
- Machine Learning (CS 229)
- Networks of Innovation
- Justice with Michael Sandel
- Nutrition Made Clear
- The Practice of Mathematics
- Joy of Mathematics
- Godel, Escher, Bach: A Mental Space Odyssey
- Mathematics from the Visual World
- Inculture(s) II

- Unsupervised Feature Learning and Deep Learning by Andrew Ng, Stanford OpenClassroom
- motivated by WithoutNotesApril11#AndrewNg
- CS 229: Machine Learning (Course handouts) Course Materials
- http://openclassroom.stanford.edu/MainFolder/CoursePage.php?course=MachineLearning

Consider PCM when clarifications are needed.

*Back to the Menu*

6.868J The Society of Mind by Marvin Minsky, MIT OpenCourseWare 2007

Mostly discussion about the definitions of intelligence.

*Back to the Menu*

List of Lecture Topics and Background Readings

- AGI versus Narrow AI by Ben Goertzel
- Approaches to Defining and Evaluating General Intelligence by Pei Wang
- http://www.cis.temple.edu/~pwang/
- http://sites.google.com/site/narswang/home/agi-introduction/agi-education
- ~1h12min mentions of projects including Soar, OSCAR, ACT-R, NARS, Polyscheme, LIDA, SNePS, AIXI, HTM then a classification of them
- see also Wikipedia:Comparison of cognitive architectures (with a link to the since deleted Wikipedia:Comparison of AGI projects)

- The Brain as a Neuroscientist Sees It 1 of 2 by Allan Combs
- The Mathematics of Universal and General Intelligence by Ben Goertzel
- A Logical Model of Intelligence 1 of 3 by Pei Wang
- The Brain as a Neuroscientist Sees It 2 of 2 by Allan Combs
- A Logical Model of Intelligence 2 of 3 by Pei Wang
- The Nature of Conscious-ness by Allan Combs
- BeingNoOne BeingNoOne
- Learning and Memory: How it Works and When it Fails by Frank Longo, Stanford Mini Med School 2010

- Introduction to Probabilistic Logic Networks by Nil Geisweiller and Ben Goertzel
- Probabilistic Logic Networks : A Comprehensive Framework for Uncertain Inference Springer 2009
- Wikipedia:Probabilistic logic network
- Probabilistic Logic Networks on OpenCog wiki
- progicnet Probabilistic logic and probabilistic networks

- A Logical Model of Intelligence 3 of 3 by Pei Wang
- Nonlinear Dynamics and the Mind by Allan Combs
- ~1h05min visualization of the semantic space as a 3D cube

- Vision and the Brain by Allan Combs
- see also previously discovered
- OpenCV (Open Source Computer Vision) is a library of programming functions for real time computer vision.
- OpenVIDIA implement computer vision algorithms running on on graphics hardware such as single or multiple graphics processing units(GPUs) using OpenGL, Cg and CUDA-C. Some samples will soon support OpenCL and Direct Compute API's also.

- see also previously discovered
- Evolvable Neural Networks by Hugo de Garis
- Managing Uncertainty with Indefinite Probabilities by Matthew Ikle
- Summary : Two key concepts
- Second-order distributions provide a theoretical framework for true uncertainty instead of presuming a particular prior
- Provides a natural mechanism measuring weight of evidence

- Summary : Two key concepts
- Program Representation for General Intelligence by Ben Goertzel and Nil Geisweiller
- Man as Machine by Joscha Bach
- Controlling Intelligent Agents in a Virtual World by Nil Geisweiller
- Understanding Motivation, Emotion and Mental Representation by Joscha Bach
- Attractor Neural Nets and Economic Attention Allocation by Joel Pitt
- Humanoid Robotics for AGI: the Nao Platform by Hugo de Garis
- Parcone (Partially Connected Neural Evolutionary)

- Extending and Customizing Virtual Worlds for AGI by Ben Goertzel
- The MicroPsi Architecture by Joscha Bach
- Imitation and Reinforcement Learning in Virtually Embodied Agents Using Program Evolution by Nil Geisweiller
- RIDE: A Simulator for Robotic Intelligence Development by HyunRyong Jung
- RIDE Robot framework Integrated Development Environment

*Back to the Menu*

Stanford CS 229: Machine Learning

- Machine Learning at Stanford OpenClassroom
- author page on the The AI Genealogy Project
- Machine Learning (CS 229), Lecture 1
- ~min32 Wikipedia:Arthur Samuel's 1959 definition
- his checker example
- Wikipedia:Tom Mitchell's 1998 definition

- ~min36 Supervised Learning
- regression
- classification

- ~min45 Learning Theory
- ~min50 Unsupervised Learning
- clustering
- ~min56 Wikipedia:Source separation and the cocktail party problem
- Wikipedia:Independent component analysis (ICA) algorithm

- 1h02min Reinforcement Learning

- ~min32 Wikipedia:Arthur Samuel's 1959 definition
- Machine Learning (CS 229), Lecture 2
- ~3min Supervised learning with "Alvin" military car driving through a 3 layers ANN (~15 years ago)
- see also Welcome DARPA Urban Challenge

- ~12min notation details
- ~26min Wikipedia:Gradient descent
- Wikipedia:Gradient
- Wikipedia:Stochastic gradient descent
- ~52min notations details
- ~58min key facts about traces and matrices derivatives

- ~1h14min Normal equations or Wikipedia:Regression analysis#General_linear_model

- ~3min Supervised learning with "Alvin" military car driving through a 3 layers ANN (~15 years ago)
- Machine Learning (CS 229), Lecture 3
- ~min3 Linear regression
- overfitting and underfitting
- ~10min parametric vs non-parametric learning algorithm
- ~12min Locally weighted regression (Wikipedia:Local regression or LOESS)

- ~28min Probabilistic interpretation

- Logistic regression
- ~49min first classification algorithm

- ~1h10min Digression Perceptron

- ~min3 Linear regression
- Machine Learning (CS 229), Lecture 4
- ~min3 Newton's method
- work on distributions (Normal, Bernouilli, Poisson, Gaussian, ...)
- ~min38 Wikipedia:Generalized linear model GLM (and how to choose an exponential distribution)
- ~min52 multinomiale

- Machine Learning (CS 229), Lecture 5
- ~1min Logistic regression example
- ~searching for a line to divide clusters in the training set
- so far, discriminative learning algorithms

- ~min4 Generative learning algorithms
- ~min7 Gaussian Digressive Algorithm (GDA)
- Digression Gaussian

- Generative & Discriminative comparison
- Naive Bayes
- Laplace Smoothing

- ~min7 Gaussian Digressive Algorithm (GDA)

- ~1min Logistic regression example
- Machine Learning (CS 229), Lecture 6
- ~1min Naive Bayes
- multivariable bernouilli
- ~7min Event models
- multinomial event model

- ~27min Neural Networks
- ~43min Support vector machines
- often considered the most efficient off-the-shelf algorithm
- Support-Vector-Machines.org by Ovidiu Ivanciuc
- SVM with polynomial kernel visualization 2007
- ~59min geometric margin

- ~1min Naive Bayes
- Machine Learning (CS 229), Lecture 7
- Optimal margin classifier
- Primal/dual optimization
- problem KKT

- SVM dual
- kernels

- Machine Learning (CS 229), Lecture 8
- Soft Margin Support Vector Machine
- Coordinate Assent Algorithm
- Sequential Minimal Optimization (SMO)

- Machine Learning (CS 229), Lecture 9
- learning theory
- bias, variance, empirical risk minimization, union bound
- Hoeffding's inequalities

- learning theory

- use
~~SVM~~(talk about result, not solution) for usage U - analysis of my browsing logs
- local (client)
- eventually limited to owns wikis network

- remote (server)
- eventually limited to own IPs

- local (client)
- classification of wiki page (even just ReadingNotes group) for automatic PmWiki:Category 'zation
- content type
- LayeredModel
- WikiBrainMapping
- other tagging proposals ToDo
- ontology targeted categorization
- aiming at scalable semantical wiki

- periodically and after large edit, reprocess and detect eventual shift

- language
- note that manual classification does not need to be done within the wiki
- an external file or a set of external files can be used to be more efficient
- even a wiki page can be used of that
- advantage : riskless, fast
- disadvantage : not integrated, not distributed, hard to handle redirections, page-only granularity

- an external file or a set of external files can be used to be more efficient

- content type
- predictions on page number, edit frequency, ...
- improve my prediction (regression/derivative) method on PBES
- live classifier
- social interactions (IM, SNS, ...)
- live wikifying
- further than spam/non-spam

- dedicated hardware
- including Electronics#FPGA
- FPGA-based MapReduce Framework for Machine Learning by Ningyi Xu, Microsoft Research, NIPS 2009

- Lyric Semiconductor Probability Processing Circuits
- AWS#MechanicalTurk for supervision
- Annotation (including Amazon's Mechanical Turk) ACL 2010 Tutorials
- The Stanford NLP (Natural Language Processing) Group Fast, Cheap, and Creative: Evaluating Translation Quality Using Amazon's Mechanical Turk EMNLP 2009
- Mechanical Turk: The Foundations A Computer Scientist in a Business School 2008

- "hotness" of a wiki page, using feature such a history of edits (size, time, frequency, ...) and network metrics (centrality, ...)

- tools
- Octave
- http://knn.mimuw.edu.pl/weboctave-project/
- with running Web interface to Octave (computation is performed on a computer at the Department of Mathematics, Warsaw University)
- doesn't have Wolfram|Alpha pre-loaded and currated datasets though

- Octave-Forge central location for the collaborative development of packages for GNU Octave
- to get serious about Octave I recommend you write a quick GreaseMonkey script to link function names in the page to Octave online doc

- http://knn.mimuw.edu.pl/weboctave-project/
- Apache Mahout Scalable machine-learning and data-mining library
- supported Algorithms
- on AWS
- Introducing Apache Mahout: Scalable, commercial-friendly machine learning for building intelligent applications by Grant Ingersoll, IBM developerWork 2009
- Intro to Apache Mahout by Grant Ingersoll, Lucid July 2010

- Octave
- Reading notes
- my previous ML resources
- Journal of Machine Learning Research (JMLR)
- CGAL: The Open Source Computational Geometry Algorithms Library, Google Tech Talks 2008
- eventually consider it with Wikipedia:Information geometry (IG) perspective

- Machine Learning at Tufts University, Computer Science Department at Tufts University
- MIT 6.867 Machine Learning Fall 2003
- ES "Apprentissage Artificiel" (machine-Learning) MINES ParisTech - MinesParis

*Back to the Menu*

Networks of Innovation by Andrew Hargadon, GSSS 2009, Santa Fe Institute 2009

- note that the talk is about innovating on sustainably mostly during the conclusion (after ~1h) and in the following Q&A
- Brown Field Investment, Investopedia
- When a company or government entity purchases or leases existing production facilities to launch a new production activity. This is one strategy used in foreign-direct investment.
- Green Field Investment, Investopedia

- min30 focus on creativity
- situation and process rather than a personality trait

- min50 connecting innovation to market a the key
- min58 on Ford and his unability to respond because of the complexity of the system he build

- min15 example of James Watt's separate condenser
- existing market
- existing demand (coal mines requiring pumps, bottleneck limiting their productivity)
- existing technology (steam engine with research being conducted)
- yet modular design to allow improvements per component

- min35 key moment was thanks to the "network that he finally accepted around him"

- min38 the fallacy of "build it and they will come", the world will not beat the path to your door despite fundamental improvements
- example of the Segway and its lack of network for marketing and distribution despite its technological advantage
- opposed to Ford and his large set of sellers before having 1 unit produced

- example of the Segway and its lack of network for marketing and distribution despite its technological advantage
- min48 sum-up in 1 sentence "
**The network**"*is*the innovation.- thus a justification for http://dashboard.seedea.org/

- min75 on the history of the Silicon Valley
- Wikipedia:Sherman Fairchild and the "Fairchild family tree"

- conclusion
- shift from the idea to the network required by this idea to make it happen
- William Gibson's famous quote "The future is already here, it is just unevenly distributed."
- your role is not to invent anything but to re-arrange the pieces existing in a new way

- Andrew Hargadon homepage
- De la gestion des brevets d'invention au pilotage de l'innovation : le cas d'un centre de recherche de haute technologie concluding on the importance of sustainable innovation
- Innovation Networks by Zann Gill on her website
- Designing Innovation Networks Modeled on Life's Origins & Evolution by Zann Gill, Google Tech Talks 2008

*Back to the Menu*

co-production of WGBH Boston and Harvard University, started in 2009

- Episode 1
- Part 1 - The Moral Side of Murder
- Part 2 - The Case for Cannibalism

*Back to the Menu*

Nutrition Made Clear by Roberta H. Anding, TTC 2009

- lot of material covered but is still rather pedagogical nothing ground breaking but everything is well explained
- (bad) habit of saying "The Science out there confirms it" without giving (eventually boring) references
- nice storytelling style experiences that she had, making it more entertaining
- against anything "miraculous" so fresh vegs and fruits from all different color and exercise seems to be my overall (too short) sum up
- no ideal diet that anybody must follow independantly of their needs and objectives
- avocating no quick fix but a real understanding of the topics Id say.

- for somebody just curious and starting to study the topic, definitly overwhelming
- the pdf is "cold" but rather complete

few practical links and at the end of the guidebook in PDF also lsits the "Credible Nutrition Websites"

- For the Exercise Sciences and Clinical Sports Medicine (ACSM)
- Centers for Disease Control and Prevention (CDC)
- Food and Drug Administration (FDA)
- United States Department of Agriculture (USDA)
- key concepts according to Wikipedia
- Basal metabolic rate (BMR)
- Thermic effect of food
- Thermogenesis and Non-exercise-associated thermogenesis (NEAT)
- Fluid balance
- Glycemic index

See also my Health#Nutrition page.

*Back to the Menu*

The Practice of Mathematics by Robert P. Langlands, Institute for Advanced Study 1999

- 16 lectures serie
- Euclid's Elements
- Euclid's Elements on Wikipedia
- The Elements of Euclid on Wiksource

- Gauss' Arithmetical Investigations
- Disquisitiones Arithmeticae on Wikipedia

*Back to the Menu*

Joy of Mathematics (first edition) by Murray H. Siegel, The Teaching Company 1999

- Concluding lecture 24 saying that
- when you control knowledge, you control the ignorants
- in the same way that in the past, Latin was the language of knowledge, now mathematics is

See also

- Joy of Mathematics last edition, new program and new teacher
- my reading notes on Princeton Companion to Mathematics
- other TTC video course on this page Mathematics from the Visual World

*Back to the Menu*

Godel, Escher, Bach: A Mental Space Odyssey by Justin Curry and Curran Kelleher, MIT OpenCourseWare 2007

- lectures
~~1 2~~~~3~~~~4~~~~5~~~~6~~(in 220K quality) - "the fundamental thing we want to answer at the end of this course : What is an 'I'. What makes something conscious from unsconscious things.How do we get particles and atoms to start talking about themselves like the way we do." lecture 3 ~42:30
- description of emergent properties makes one wonder if it's an epiphenomenon or if we just "create" an higher being based on the similarity of behaviors
- not using a biological boundary but a behavioral boundary of the organism

- is the concept of "self" regenerated consantly in the same way that biological cells are and that we are not the same person than we were weeks ago?
- if so, can we find artefacts and social means to build and support resiliance of our self? i.e. ss our social context helping us to maintain who we are?

- lectures

See also my notes on I am a Strange Loop

*Back to the Menu*

Mathematics from the Visual World by Michael Starbird, Course No. 1447, The Teaching Company 2009

Why am I watching this.

Struture, actual essence, objective review, ...

01 - Seing with Our Eyes, Seing with Our Minds

- historical introduction
- usage of geometry, going from the abstract back to the real
- mathematics as the "exploration of self-consistent systems"
- different components including theorems, axioms, premisses, proofs, ...
- great proofs display underlying connections

- Euclides elements
- definitions/axioms/commons notions
- point, line, angle, ...

- postulates
- parallel lines, ...

- definitions/axioms/commons notions
- proving that the sum of the angles of a triangle is 180 degrees
- using the pallelel lines with its crossing line transversal postulate

- proving that the angles of an isocele triangle are equal
- flipping through its symmetry axis

- "There is no royal road to Geometry." as Euclid replied to Ptolemy in Elements (ed. Thomas L. Heath)

02 - Congruence, Similarity, and Pythagoras

- classical mathematical problem to study the meaning of "equal", sameness
- congruence
- one figure can be moved over the other exactly
- SSS: side side side congruence theorem for triangles
- usage of this property of triangle in architecture (i.e. rigidity as a key required component for bridges, ...)

- SAS: side angle side
- ASA : angle side angle
~~SSA: side side angle~~is not a proper theorem- more examples with different shapes

- similarity
- same shapes but potentialy of different size
- triangle
- 2 same angles
- same expansion factors (3 lengths as ratios of the corresponding lengths)
- 2 sides of the same expansions factors and a same angle (SAS+proportional sides)

- more examples with different shapes
- Eiffel tower measurments using Michael Starbird's shadow and similarity property

- basis of trigonometry
- proofs of the pythagorian theorem using congruence and similarity
- for any right triangle ABC, AB
^{2}+BC^{2}=AC^{2} - Leonardo Da Vinci, using area and rotations 2*ABC+AC
^{2}=2*ABC+AB^{2}+BC^{2}- consequently, through rotation it uses congruence

- forming similar triangles through the hypothenuse

- for any right triangle ABC, AB

03 - The Circle

- shape famous accross culture, probably the most popular symbol
- three (non-aligned) points determinate a circle
- relationship between the circumference and the diameter
- constant Pi
- C=D*Pi=2*Pi*r

- properties
- angle
- chord
- tangent

- construction of an Apollonian circle

04 - Centers of Triangles

- median, from a vertex to the middle of a side
- examples of balancing on a ruler

- centroid, where the medians meet
- incenter, where the bisectors meet, center of the inner circle
- circumcenter, center of the outer circle
- altitudes, from a vertex to a side, orthogonally
- orthocenter, where the altitudes meet
- barycentre in french?
- Euler line, containing the circumcenter, the centroid and the orthocenter

05 - Surprising Complexity of Simple Triangles

- nine points cirle
- proof strategy of similarity, parallelness and right angle triangle within a circle

- Napoleon's theorem
- proof strategy of teslating the plane

- Morley's miracle
- proof strategy of similarity

06 - Clever Constructions

- constructing with straight edge and compass
- regular polygons
- bisections
- copying an angle
- golden rectangle

- constructible distance
- using a unit length, 1
- sum
- multiplication
- quotient
- square root

07 - Impossible Geometry - Squaring the Circle

- doubling the cube, Delian question
- Galois theory

- trisection of an angle
- squaring of a circle

08 - Classic Conics

- conic sections
- ?

09 - Amazing Areas

- area of classical shapes
- deducing for a circle then an ellipsde
- Mamikon Mnatsakanian approach
- Visual Calculus on Wikipedia
- his CV at CALTech

Inculture 2 - Une autre histoire de l'Ã©ducation, Franck Le Page, 2009

- rappel de la "1ere annee" de la bulle d'air chaud representant l'ascension sociale
- ~10min revolution francaise -> societe egalitaire donc utilisation de l'ecole
- 5 plans d'education
- Condorcet (retenue)
- pas une ecole du savoir minimum
- ecole du savoirmaximum, chacun doit pouvoir aller aussi loin qu'il lui est possible
- risque de faire une societe ou les inegalites seront bases sur les savoirs

- Condorsay lui-meme a confirme le risque de la creation d'une elite mais a affirme que cette elite se met au service de la nation

- Mirabeau
- Talerant (?)
- Le peletier de Saint-Fargeau, defendu par Robespierre
- Lepeletier de Saint-Fargeau par James Guillaume, Institut National de Recherche Pédagogique (INRP)
- proposer de frainer les plus rapides pour avancer ensemble
- pedagogie Fresnay en a herite

- ... dernier inconnu

- Condorcet (retenue)
- ecole qui frabrique de l'inegalite
- notation individuelle
- constante macabre

- 5 plans d'education
- ~28min l'egalite des chances c'est que la Republique garantie que
- "pour le lievre comme pour la tortue, la ligne de depart est la meme"

- resume :
**lecon de deconstruction de la plaisanterie de l'ascencion sociale par la culture** - ~33min "ces micro-competences sont toutes des competences a obeir a ce qui est demande." cf Chomsky
- Bernard Defrance, philosophe, "L'ecole consiste a repeter ce que dit le maitre."
- Bernard Charlot qui a travaille sur le rapport au savoir
- sa page de membre associé du Laboratoire Innovation Formation Education (LIFE)

- Aristotle : "Those that know, do. Those that understand, teach."
- site de l’Appel pour une école démocratique propose par http://fr.wikipedia.org/wiki/Nico_HirttNico Hirtt
- Le blog de Nico Hirtt: Une école, ça change le monde sur Mediapart
- Les trois axes de la marchandisation scolaire, articile fondateur par Nico Hirtt, 2001

- ~2h15min schema base sur les donnees INSEE
- indicateur du niveau culturel global de la population francaise vs. besoin de entreprises
- historique et evolution
- phenomene de dequalification a l'embauche, surqualification a l'emploi juste apres la guerre
- inversement vers 1973
- phenomene de surqualification a l'embauche, dequalification a l'emploi
- engage pour faire moins que ce que l'on sait faire
- organisation du travail dans l'entreprise, propriete prive des moyens de production
- "lutte des places"

- ~2h23min conclusion
- instauration d'un systeme d'education reellement egalitaire (pas de competition, pas de carte carte scolaire, college unique, acces a l'universite pour tous, pas de grandes ecoles, ...)

- fonction instruction (independance intelectuelle) / d'education (preparation professionnel ) / formation pro

- la presentation de Lessig sur le status du salarier
- L'école idéale, ZDF, 2008
- initiallement Die ideale Schule de 2h

- La République des idées
- L’école des chances by François Dubet
- L'élitisme républicain by Christian Baudelot and Roger Establet

- http://en.wikipedia.org/wiki/Grandes_ecoles#Origins

*Back to the Menu*

- mimic the structure from ReadingNotes
- add previous notes
- only multi-emissions content from WithoutNotes (mainly courses)

*Back to the Menu*