19 Better Neural Network Training; Convolutional Neural Networks

Size: px
Start display at page:

Download "19 Better Neural Network Training; Convolutional Neural Networks"

Transcription

1 108 Jnathan Richard Shewchuk 19 Better Neural Netwrk Training; Cnvlutinal Neural Netwrks [I m ging t talk abut a bunch f heuristics that make gradient descent faster, r make it find better lcal minima, r prevent it frm verfitting. I suggest yu implement vanilla stchastic backprp first, and experiment with the ther heuristics nly after yu get that wrking.] Heuristics fr Faster Training [A big disadvantage f neural nets is that they take a lng, lng time t train cmpared t ther classificatin methds we ve studied. Here are sme ways t speed them up. Unfrtunately, yu usually have t experiment with techniques and hyperparameters t find which nes will help with yur particular applicatin.] (1) (5) abve. [Fix the unit saturatin prblem.] Stchastic gradient descent: faster than batch n large, redundant data sets. [Whereas batch gradient descent walks dwnhill n ne cst functin, stchastic descent takes a very shrt step dwnhill n ne pint s lss functin and then anther shrt step n anther pint s lss functin. The cst functin is the sum f the lss functins ver all the sample pints, s ne batch step behaves similarly t n stchastic steps and takes the same amunt f time. But if yu have many di erent examples f the number 9, they cntain much redundant infrmatin, and stchastic gradient descent learns the redundant infrmatin mre quickly.] y batchvsstch.pdf (LeCun et al., E cient BackPrp ) [Left: a simple neural net with nly three weights, and its 2D training data. Center: batch gradient descent makes nly a little prgress each epch. (Epchs alternate between red and blue.) Right: stchastic descent decreases the errr much faster than batch descent.] One epch presents every training example nce. Training usually takes many epchs, but if sample is huge [and carries lts f redundant infrmatin] it can take less than ne.

2 Better Neural Netwrk Training; Cnvlutinal Neural Netwrks 109 Nrmalizing the data. Center each feature s mean is zer. Then scale each feature s variance 1. [The first step seems t make it easier fr hidden units t get int a gd perating regin f the sigmid r ReLU. The secnd step makes the bjective functin better cnditined, s gradient descent cnverges faster.] nrmalize.jpg [A 2D example f data nrmalizatin.] 6 x x illcnditin.pdf [Skewed data ften leads t an bjective functin with an ill-cnditined (highly eccentric) Hessian. Gradient descent in these functins can be painfully slw, as this figure shws. Nrmalizing helps by reducing the eccentricity. Whitening reduces the eccentricity even mre, but it s much mre expensive. Anther thing that helps with elngated trughs like this is mmentum, which we ll discuss shrtly. It eventually gets us mving fast alng the lng axis.]

3 110 Jnathan Richard Shewchuk Centering the hidden units helps t. Replace sigmids with tanh = e e e +e = 2 s(2 ) 1. [This functin ranges frm 1 t 1 instead f frm 0 t 1.] [If yu use tanh units, dn t frget that yu als need t change backprp t replace s 0 with the derivative f tanh. Als, gd utput target values change t rughly 0.7 and 0.7.] Di erent learning rate fr each layer f weights. Earlier layers have smaller gradients, need larger learning rate. curvaturelayers.pdf [In this illustratin, the inputs are at the bttm, and the utputs at the tp. The derivatives tend t be smaller at the earlier layers.] Emphasizing schemes. [Neural netwrks learn the mst redundant examples quickly, and the mst rare examples slwly. S we try t emphasize the uncmmn examples.] Present examples frm rare classes mre ften, r w/bigger. Same fr misclassified examples. [I have t warn yu that emphasizing schemes can backfire n bad utliers.] Secnd-rder ptimizatin. [Unfrtunately, Newtn s methd is cmpletely impractical, because the Hessian is t large and expensive t cmpute. There have been a lt f attempts t incrprate curvature infrmatin int neural net learning in cheaper ways, but nne f them are ppular yet.] Nnlinear cnjugate gradient: wrks well fr small nets + small data + regressin. Batch descent nly!! T slw with redundant data. Stchastic Levenberg Marquardt: apprximates a diagnal Hessian. [The authrs claim cnvergence is typically three times faster than well-tuned stchastic gradient descent. The algrithm is cmplicated.]

4 Better Neural Netwrk Training; Cnvlutinal Neural Netwrks 111 Heuristics fr Aviding Bad Lcal Minima (1) (5) abve. [Fix the unit saturatin prblem.] Stchastic gradient descent. Randm mtin gets yu ut f shallw lcal minima. [Even if yu re at a lcal minimum f the cst functin, each sample pint s lss functin will be trying t push yu in a di erent directin. Stchastic gradient descent lks like a randm walk r Brwnian mtin, which can shake yu ut f a shallw minimum.] Mmentum. Gradient descent changes velcity slwly. Carries us thrugh shallw lcal minima t deeper nes. w rw repeat w w + w w rw + w Gd fr bth batch & stchastic. Chse hyperparameter <1. [Think f w as the velcity. The hyperparameter specifies hw much mmentum persists frm iteratin t iteratin.] [I ve seen cnflicting advice n. Sme researchers set it t 0.9; sme set it clse t zer. Ge Hintn suggests starting at 0.5 and slwly increasing it t 0.9 r higher as the gradients get small.] [If is large, yu shuld usually chse smaller t cmpensate, and yu might als want t change the first line s the initial velcity is larger.] [A prblem with mmentum is that nce it gets clse t a gd minimum, it scillates arund the minimum. But it s mre likely t get clse t a gd minimum in the first place.] Heuristics t Avid Overfitting Ensemble f neural nets. Randm initial weights + bagging. [We saw hw well ensemble learning wrks fr decisin trees. It wrks well fr neural nets t. The cmbinatin f randm initial weights and bagging helps ensure that each neural net cmes ut di erently. Obviusly, ensembles f neural nets are slw.] `2 regularizatin, aka weight decay. Add w 0 2 t the cst/lss fn, where w 0 is vectr f all weights except biases. [w 0 includes all the weights in matrices V and W except the last clumn f each, rewritten as a vectr.] [We d this fr the same reasn we d it in ridge regressin: penalizing large weights reduces verfitting by reducing the variance f the methd.] i has extra term 2 w i Weight decays by factr 1 2 if nt reinfrced by training.

5 112 Jnathan Richard Shewchuk Neural Netwrk - 10 Units, N Weight Decay Neural Netwrk - 10 Units, Weight Decay= Training Errr: Test Errr: Bayes Errr: Training Errr: Test Errr: Bayes Errr: weightdecay.pdf, weightdecayn.pdf (ESL, Figure 11.4) Write sftmax + crssentrpy lss. [Examples f 2D classificatin withut (left) and with (right) weight decay. Observe that the secnd example better apprximates the Bayes ptimal bundary (dashed purple curve).] Drput emulates an ensemble in ne netwrk. drput.pdf [During training, we temprarily disable a randm subset f the units, alng with all the edges in and ut f thse units. It seems t wrk well t disable each hidden unit with prbability 0.5, and t disable input units with a smaller prbability. We d stchastic gradient descent and we frequently change which randm subset f units is disabled. The authrs claim that their methd gives even better generalizatin than `2 regularizatin. It gives sme f the advantages f an ensemble, but it s faster t train.] Fewer hidden units. [The number f hidden units is a hyperparameter yu can use t adjust the bias-variance trade. If there s t few, yu can t learn well, but if there s t many, yu tend t verfit. `2 regularizatin and drput make it safer t have t many hidden units, s it s less critical t find just the right number.]

6 Better Neural Netwrk Training; Cnvlutinal Neural Netwrks 113 CONVOLUTIONAL NEURAL NETWORKS (CnvNets; CNNs) [Cnvlutinal neural nets have caused a big resurgence f interest in neural nets in the last few years. Often yu ll hear the buzzwrd deep learning, which refers t neural nets with many layers. Many f the mst successful deep netwrks have been cnvlutinal.] Visin: inputs are large images image = 40,000 pixels. If we cnnect them all t 40,000 hidden units! 1.6 billin cnnectins. Neural nets are ften verparametrized: t many weights, t little data. [As a rule f thumb, yu shuld have mre data than yu have weights t estimate. If yu dn t fllw that rule, yu usually verfit. With images, it s impssible t get enugh data t crrectly train billins f weights. We culd shrink the images, but then we re thrwing away useful data. Anther prblem with having billins f weights is that the netwrk becmes very slw t train r even t use.] [Researchers have addressed these prblems by taking inspiratin frm the neurlgy f the visual system. Remember that early in the semester, I tld yu that yu can get better perfrmance n the handwriting recgnitin task by using edge detectrs. Edge detectrs have tw interesting prperties. First, each edge detectr lks at just ne small part f the image. Secnd, the edge detectin cmputatin is the same n matter which part f the image yu apply it t. S let s apply these tw prperties t neural net design.] CnvNet ideas: (1) Lcal cnnectivity: A hidden unit (in early layer) cnnects nly t small patch f units in previus layer. [This imprves the verparametrizatin prblem, and speeds up bth the frward pass and the training cnsiderably.] (2) Shared weights: Grups f hidden units share same set f input weights, called a mask aka filter aka kernel. [N relatin t the kernels f Lecture 14.] We learn several masks. [Each mask perates n every patch f image.] Masks patches = hidden units in first hidden layer. If ne patch learns t detect edges, every patch has an edge detectr. [Because the mask that detects edges is applied t every patch.] CnvNets explit repeated structure in images, audi. Cnvlutin: the same linear transfrmatin applied t di erent parts f the input by shifting. [Shared weights imprve the verparametrizatin prblem even mre, because shared weights means fewer weights. It s a kind f regularizatin.] [But shared weights have anther big advantage. Suppse that gradient descent starts t develp an edge detectr. That edge detectr is being trained n every part f every image, nt just n ne spt. And that s gd, because edges appear at di erent lcatins in di erent images. The lcatin n lnger matters; the edge detectr can learn frm edges in every part f the image.] [In a neural net, yu can think f hidden units as features that we learn, as ppsed t features that yu cde up yurself. Cnvlutinal neural nets take them t the next level by learning features frm multiple patches simultaneusly and then applying thse features everywhere, nt just in the patches where they were riginally learned.] [By the way, althugh lcal cnnectivity is inspired by ur visual systems, shared weights bviusly dn t happen in bilgy.]

7 114 Jnathan Richard Shewchuk [Shw slides n cmputing in the visual crtex and CnvNets (cnn.pdf), available frm the CS 189 web page. Srry, readers, there are t many images t include here. The narratin is belw.] [Neurlgists can stick needles int individual neurns in animal brains. After a few hurs the neurn dies, but until then they can recrd its actin ptentials. In this way, bilgists quickly learned hw sme f the neurns in the retina, called retinal ganglin cells, respnd t light. They have interesting receptive fields, illustrated in the slides, which shw that each ganglin cell receives excitatry stimulatin frm receptrs in a small patch f the retina but inhibitry stimulatin frm ther receptrs arund it.] [The signals frm these cells prpagate t the V1 visual crtex in the ccipital lbe at the back f yur skull. The V1 cells prved much harder t understand. David Hubel and Trsten Wiesel f the Jhns Hpkins University put prbes int the V1 visual crtex f cats, but they had a very hard time getting any neurns t fire there. Hwever, a lucky accident unlcked the secret and ultimately wn them the 1981 Nbel Prize in Physilgy.] [Shw vide HubelWiesel.mp4, taken frm YuTube: ] [The glass slide happened t be at the particular rientatin the neurn was sensitive t. The neurn desn t respnd t ther rientatins; just that ne. S they were pretty lucky t catch that.] [The simple cells act as line detectrs and/r edge detectrs by taking a linear cmbinatin f inputs frm retinal ganglin cells.] [The cmplex cells act as lcatin-independent line detectrs by taking inputs frm many simple cells, which are lcatin dependent.] [Later researchers shwed that lcal cnnectivity runs thrugh the V1 crtex by prjecting certain images nt the retina and using radiactive tracers in the crtex t mark which neurns had been firing. Thse images shw that the neural mapping frm the retina t V1 is retinatpic, i.e., lcality preserving. This is a big part f the inspiratin fr cnvlutinal neural netwrks!] [Unfrtunately, as we g deeper int the visual system, layers V2 and V3 and s n, we knw less and less abut what prcessing the visual crtex des.] [CnvNets were first ppularized by the success f Yann LeCun s LeNet 5 handwritten digit recgnitin sftware. LeNet 5 has six hidden layers! Layers 1 and 3 are cnvlutinal layers in which grups f units share weights. Layers 2 and 4 are pling layers that make the image smaller. These are just hardcded max-functins with n weights and nthing t train. Layers 5 and 6 are just regular layers f hidden units with n shared weights. A great deal f experimentatin went int figuring ut the number f layers and their sizes. At its peak, LeNet 5 was respnsible fr reading the zip cdes n 10% f US Mail. Anther Yann LeCun system was deplyed in ATMs and check reading machines and was reading 10 t 20% f all the checks in the US by the late 90 s.] [Shw Yann LeCun s vide LeNet5.mv, illustrating LeNet 5.]

8 Better Neural Netwrk Training; Cnvlutinal Neural Netwrks 115 LeNet5.png Architecture f LeNet5. [When CnvNets were first applied t image analysis, researchers fund that sme f the learned masks are edge detectrs r line detectrs, similar t the nes that Hubel and Wiesel discvered! This created a lt f excitement in bth the cmputer learning cmmunity and the neurscience cmmunity. The fact that a neural net can naturally learn the same features as the mammalian visual crtex is impressive.] [I tld yu tw lectures ag that neural nets research was ppular in the 60 s, but the 1969 bk Perceptrns killed interest in them thrughut the 70 s. They came back in the 80 s, but interest was partly killed a secnd time in the 00 s by... guess what? By supprt vectr machines. SVMs wrk well fr a lt f tasks, they re much faster t train, and they mre r less have nly ne hyperparameter, whereas neural nets take a lt f wrk t tune.] [Neural nets are nw in their third wave f ppularity. The single biggest factr in bringing them back is prbably big data. Thanks t the internet, we nw have abslutely huge cllectins f images t train neural nets with, and researchers have discvered that neural nets ften give better perfrmance than cmpeting algrithms when yu have huge amunts f data t train them with. In particular, cnvlutinal neural nets are nw learning better features than hand-tuned features. That s a recent change.] [One event that brught attentin back t neural nets was the ImageNet Image Classificatin Challenge in The winner f that cmpetitin was a neural net, and it wn by a huge margin, abut 10%. It s called AlexNet, and it s surprisingly similarly t LeNet 5, in terms f hw its layers are structured. Hwever, there are sme new innvatins that led t their prize-winning perfrmance, in additin t the fact that the training set had 1.4 millin images: they used ReLUs, GPUs fr training, and drput.] alexnet.pdf Architecture f AlexNet.

IAML: Support Vector Machines

IAML: Support Vector Machines 1 / 22 IAML: Supprt Vectr Machines Charles Suttn and Victr Lavrenk Schl f Infrmatics Semester 1 2 / 22 Outline Separating hyperplane with maimum margin Nn-separable training data Epanding the input int

More information

x 1 Outline IAML: Logistic Regression Decision Boundaries Example Data

x 1 Outline IAML: Logistic Regression Decision Boundaries Example Data Outline IAML: Lgistic Regressin Charles Suttn and Victr Lavrenk Schl f Infrmatics Semester Lgistic functin Lgistic regressin Learning lgistic regressin Optimizatin The pwer f nn-linear basis functins Least-squares

More information

In SMV I. IAML: Support Vector Machines II. This Time. The SVM optimization problem. We saw:

In SMV I. IAML: Support Vector Machines II. This Time. The SVM optimization problem. We saw: In SMV I IAML: Supprt Vectr Machines II Nigel Gddard Schl f Infrmatics Semester 1 We sa: Ma margin trick Gemetry f the margin and h t cmpute it Finding the ma margin hyperplane using a cnstrained ptimizatin

More information

Chapter 7. Neural Networks

Chapter 7. Neural Networks Chapter 7. Neural Netwrks Wei Pan Divisin f Bistatistics, Schl f Public Health, University f Minnesta, Minneaplis, MN 55455 Email: weip@bistat.umn.edu PubH 7475/8475 c Wei Pan Intrductin Chapter 11. nly

More information

Slide04 (supplemental) Haykin Chapter 4 (both 2nd and 3rd ed): Multi-Layer Perceptrons

Slide04 (supplemental) Haykin Chapter 4 (both 2nd and 3rd ed): Multi-Layer Perceptrons Slide04 supplemental) Haykin Chapter 4 bth 2nd and 3rd ed): Multi-Layer Perceptrns CPSC 636-600 Instructr: Ynsuck Che Heuristic fr Making Backprp Perfrm Better 1. Sequential vs. batch update: fr large

More information

COMP 551 Applied Machine Learning Lecture 11: Support Vector Machines

COMP 551 Applied Machine Learning Lecture 11: Support Vector Machines COMP 551 Applied Machine Learning Lecture 11: Supprt Vectr Machines Instructr: (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted fr this curse

More information

Admin. MDP Search Trees. Optimal Quantities. Reinforcement Learning

Admin. MDP Search Trees. Optimal Quantities. Reinforcement Learning Admin Reinfrcement Learning Cntent adapted frm Berkeley CS188 MDP Search Trees Each MDP state prjects an expectimax-like search tree Optimal Quantities The value (utility) f a state s: V*(s) = expected

More information

COMP 551 Applied Machine Learning Lecture 4: Linear classification

COMP 551 Applied Machine Learning Lecture 4: Linear classification COMP 551 Applied Machine Learning Lecture 4: Linear classificatin Instructr: Jelle Pineau (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted

More information

Pattern Recognition 2014 Support Vector Machines

Pattern Recognition 2014 Support Vector Machines Pattern Recgnitin 2014 Supprt Vectr Machines Ad Feelders Universiteit Utrecht Ad Feelders ( Universiteit Utrecht ) Pattern Recgnitin 1 / 55 Overview 1 Separable Case 2 Kernel Functins 3 Allwing Errrs (Sft

More information

Activity Guide Loops and Random Numbers

Activity Guide Loops and Random Numbers Unit 3 Lessn 7 Name(s) Perid Date Activity Guide Lps and Randm Numbers CS Cntent Lps are a relatively straightfrward idea in prgramming - yu want a certain chunk f cde t run repeatedly - but it takes a

More information

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification COMP 551 Applied Machine Learning Lecture 5: Generative mdels fr linear classificatin Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Jelle Pineau Class web page: www.cs.mcgill.ca/~hvanh2/cmp551

More information

Physics 212. Lecture 12. Today's Concept: Magnetic Force on moving charges. Physics 212 Lecture 12, Slide 1

Physics 212. Lecture 12. Today's Concept: Magnetic Force on moving charges. Physics 212 Lecture 12, Slide 1 Physics 1 Lecture 1 Tday's Cncept: Magnetic Frce n mving charges F qv Physics 1 Lecture 1, Slide 1 Music Wh is the Artist? A) The Meters ) The Neville rthers C) Trmbne Shrty D) Michael Franti E) Radiatrs

More information

SPH3U1 Lesson 06 Kinematics

SPH3U1 Lesson 06 Kinematics PROJECTILE MOTION LEARNING GOALS Students will: Describe the mtin f an bject thrwn at arbitrary angles thrugh the air. Describe the hrizntal and vertical mtins f a prjectile. Slve prjectile mtin prblems.

More information

AP Physics Kinematic Wrap Up

AP Physics Kinematic Wrap Up AP Physics Kinematic Wrap Up S what d yu need t knw abut this mtin in tw-dimensin stuff t get a gd scre n the ld AP Physics Test? First ff, here are the equatins that yu ll have t wrk with: v v at x x

More information

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must M.E. Aggune, M.J. Dambrg, M.A. El-Sharkawi, R.J. Marks II and L.E. Atlas, "Dynamic and static security assessment f pwer systems using artificial neural netwrks", Prceedings f the NSF Wrkshp n Applicatins

More information

Physics 2010 Motion with Constant Acceleration Experiment 1

Physics 2010 Motion with Constant Acceleration Experiment 1 . Physics 00 Mtin with Cnstant Acceleratin Experiment In this lab, we will study the mtin f a glider as it accelerates dwnhill n a tilted air track. The glider is supprted ver the air track by a cushin

More information

ENSC Discrete Time Systems. Project Outline. Semester

ENSC Discrete Time Systems. Project Outline. Semester ENSC 49 - iscrete Time Systems Prject Outline Semester 006-1. Objectives The gal f the prject is t design a channel fading simulatr. Upn successful cmpletin f the prject, yu will reinfrce yur understanding

More information

Artificial Neural Networks MLP, Backpropagation

Artificial Neural Networks MLP, Backpropagation Artificial Neural Netwrks MLP, Backprpagatin 01001110 01100101 01110101 01110010 01101111 01101110 01101111 01110110 01100001 00100000 01110011 01101011 01110101 01110000 01101001 01101110 01100001 00100000

More information

COMP9444 Neural Networks and Deep Learning 3. Backpropagation

COMP9444 Neural Networks and Deep Learning 3. Backpropagation COMP9444 Neural Netwrks and Deep Learning 3. Backprpagatin Tetbk, Sectins 4.3, 5.2, 6.5.2 COMP9444 17s2 Backprpagatin 1 Outline Supervised Learning Ockham s Razr (5.2) Multi-Layer Netwrks Gradient Descent

More information

Resampling Methods. Cross-validation, Bootstrapping. Marek Petrik 2/21/2017

Resampling Methods. Cross-validation, Bootstrapping. Marek Petrik 2/21/2017 Resampling Methds Crss-validatin, Btstrapping Marek Petrik 2/21/2017 Sme f the figures in this presentatin are taken frm An Intrductin t Statistical Learning, with applicatins in R (Springer, 2013) with

More information

INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2

INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2 Intr t the Physilgy f Perceptin 1 Perceptin (PSY 4204) Christine L. Ruva, Ph.D. INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2 RECEPTORS & NEURAL PROCESSING Theme f Chapter: We d nt just perceive

More information

The Law of Total Probability, Bayes Rule, and Random Variables (Oh My!)

The Law of Total Probability, Bayes Rule, and Random Variables (Oh My!) The Law f Ttal Prbability, Bayes Rule, and Randm Variables (Oh My!) Administrivia Hmewrk 2 is psted and is due tw Friday s frm nw If yu didn t start early last time, please d s this time. Gd Milestnes:

More information

k-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels

k-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels Mtivating Example Memry-Based Learning Instance-Based Learning K-earest eighbr Inductive Assumptin Similar inputs map t similar utputs If nt true => learning is impssible If true => learning reduces t

More information

Chapter 3: Cluster Analysis

Chapter 3: Cluster Analysis Chapter 3: Cluster Analysis } 3.1 Basic Cncepts f Clustering 3.1.1 Cluster Analysis 3.1. Clustering Categries } 3. Partitining Methds 3..1 The principle 3.. K-Means Methd 3..3 K-Medids Methd 3..4 CLARA

More information

T Algorithmic methods for data mining. Slide set 6: dimensionality reduction

T Algorithmic methods for data mining. Slide set 6: dimensionality reduction T-61.5060 Algrithmic methds fr data mining Slide set 6: dimensinality reductin reading assignment LRU bk: 11.1 11.3 PCA tutrial in mycurses (ptinal) ptinal: An Elementary Prf f a Therem f Jhnsn and Lindenstrauss,

More information

Enhancing Performance of MLP/RBF Neural Classifiers via an Multivariate Data Distribution Scheme

Enhancing Performance of MLP/RBF Neural Classifiers via an Multivariate Data Distribution Scheme Enhancing Perfrmance f / Neural Classifiers via an Multivariate Data Distributin Scheme Halis Altun, Gökhan Gelen Nigde University, Electrical and Electrnics Engineering Department Nigde, Turkey haltun@nigde.edu.tr

More information

Support-Vector Machines

Support-Vector Machines Supprt-Vectr Machines Intrductin Supprt vectr machine is a linear machine with sme very nice prperties. Haykin chapter 6. See Alpaydin chapter 13 fr similar cntent. Nte: Part f this lecture drew material

More information

, which yields. where z1. and z2

, which yields. where z1. and z2 The Gaussian r Nrmal PDF, Page 1 The Gaussian r Nrmal Prbability Density Functin Authr: Jhn M Cimbala, Penn State University Latest revisin: 11 September 13 The Gaussian r Nrmal Prbability Density Functin

More information

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d)

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d) COMP 551 Applied Machine Learning Lecture 9: Supprt Vectr Machines (cnt d) Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Class web page: www.cs.mcgill.ca/~hvanh2/cmp551 Unless therwise

More information

Lesson Plan. Recode: They will do a graphic organizer to sequence the steps of scientific method.

Lesson Plan. Recode: They will do a graphic organizer to sequence the steps of scientific method. Lessn Plan Reach: Ask the students if they ever ppped a bag f micrwave ppcrn and nticed hw many kernels were unppped at the bttm f the bag which made yu wnder if ther brands pp better than the ne yu are

More information

This section is primarily focused on tools to aid us in finding roots/zeros/ -intercepts of polynomials. Essentially, our focus turns to solving.

This section is primarily focused on tools to aid us in finding roots/zeros/ -intercepts of polynomials. Essentially, our focus turns to solving. Sectin 3.2: Many f yu WILL need t watch the crrespnding vides fr this sectin n MyOpenMath! This sectin is primarily fcused n tls t aid us in finding rts/zers/ -intercepts f plynmials. Essentially, ur fcus

More information

[COLLEGE ALGEBRA EXAM I REVIEW TOPICS] ( u s e t h i s t o m a k e s u r e y o u a r e r e a d y )

[COLLEGE ALGEBRA EXAM I REVIEW TOPICS] ( u s e t h i s t o m a k e s u r e y o u a r e r e a d y ) (Abut the final) [COLLEGE ALGEBRA EXAM I REVIEW TOPICS] ( u s e t h i s t m a k e s u r e y u a r e r e a d y ) The department writes the final exam s I dn't really knw what's n it and I can't very well

More information

Getting Involved O. Responsibilities of a Member. People Are Depending On You. Participation Is Important. Think It Through

Getting Involved O. Responsibilities of a Member. People Are Depending On You. Participation Is Important. Think It Through f Getting Invlved O Literature Circles can be fun. It is exciting t be part f a grup that shares smething. S get invlved, read, think, and talk abut bks! Respnsibilities f a Member Remember a Literature

More information

The blessing of dimensionality for kernel methods

The blessing of dimensionality for kernel methods fr kernel methds Building classifiers in high dimensinal space Pierre Dupnt Pierre.Dupnt@ucluvain.be Classifiers define decisin surfaces in sme feature space where the data is either initially represented

More information

Day 7: Optimization, regularization for NN

Day 7: Optimization, regularization for NN Day 7: Optimizatin, regularizatin fr NN Intrductin t Machine Learning Summer Schl June 18, 2018 - June 29, 2018, Chicag Instructr: Suriya Gunasekar, TTI Chicag 26 June 2018 Day 7: Tricks and tls Day 7:

More information

Physical Layer: Outline

Physical Layer: Outline 18-: Intrductin t Telecmmunicatin Netwrks Lectures : Physical Layer Peter Steenkiste Spring 01 www.cs.cmu.edu/~prs/nets-ece Physical Layer: Outline Digital Representatin f Infrmatin Characterizatin f Cmmunicatin

More information

Part 3 Introduction to statistical classification techniques

Part 3 Introduction to statistical classification techniques Part 3 Intrductin t statistical classificatin techniques Machine Learning, Part 3, March 07 Fabi Rli Preamble ØIn Part we have seen that if we knw: Psterir prbabilities P(ω i / ) Or the equivalent terms

More information

I. Analytical Potential and Field of a Uniform Rod. V E d. The definition of electric potential difference is

I. Analytical Potential and Field of a Uniform Rod. V E d. The definition of electric potential difference is Length L>>a,b,c Phys 232 Lab 4 Ch 17 Electric Ptential Difference Materials: whitebards & pens, cmputers with VPythn, pwer supply & cables, multimeter, crkbard, thumbtacks, individual prbes and jined prbes,

More information

BASD HIGH SCHOOL FORMAL LAB REPORT

BASD HIGH SCHOOL FORMAL LAB REPORT BASD HIGH SCHOOL FORMAL LAB REPORT *WARNING: After an explanatin f what t include in each sectin, there is an example f hw the sectin might lk using a sample experiment Keep in mind, the sample lab used

More information

Differentiation Applications 1: Related Rates

Differentiation Applications 1: Related Rates Differentiatin Applicatins 1: Related Rates 151 Differentiatin Applicatins 1: Related Rates Mdel 1: Sliding Ladder 10 ladder y 10 ladder 10 ladder A 10 ft ladder is leaning against a wall when the bttm

More information

What is Statistical Learning?

What is Statistical Learning? What is Statistical Learning? Sales 5 10 15 20 25 Sales 5 10 15 20 25 Sales 5 10 15 20 25 0 50 100 200 300 TV 0 10 20 30 40 50 Radi 0 20 40 60 80 100 Newspaper Shwn are Sales vs TV, Radi and Newspaper,

More information

Chapter 11: Neural Networks

Chapter 11: Neural Networks Chapter 11: Neural Netwrks DD3364 December 16, 2012 Prjectin Pursuit Regressin Prjectin Pursuit Regressin mdel: Prjectin Pursuit Regressin f(x) = M g m (wmx) t i=1 where X R p and have targets Y R. Additive

More information

https://goo.gl/eaqvfo SUMMER REV: Half-Life DUE DATE: JULY 2 nd

https://goo.gl/eaqvfo SUMMER REV: Half-Life DUE DATE: JULY 2 nd NAME: DUE DATE: JULY 2 nd AP Chemistry SUMMER REV: Half-Life Why? Every radiistpe has a characteristic rate f decay measured by its half-life. Half-lives can be as shrt as a fractin f a secnd r as lng

More information

Distributions, spatial statistics and a Bayesian perspective

Distributions, spatial statistics and a Bayesian perspective Distributins, spatial statistics and a Bayesian perspective Dug Nychka Natinal Center fr Atmspheric Research Distributins and densities Cnditinal distributins and Bayes Thm Bivariate nrmal Spatial statistics

More information

CESAR Science Case The differential rotation of the Sun and its Chromosphere. Introduction. Material that is necessary during the laboratory

CESAR Science Case The differential rotation of the Sun and its Chromosphere. Introduction. Material that is necessary during the laboratory Teacher s guide CESAR Science Case The differential rtatin f the Sun and its Chrmsphere Material that is necessary during the labratry CESAR Astrnmical wrd list CESAR Bklet CESAR Frmula sheet CESAR Student

More information

AP Statistics Notes Unit Two: The Normal Distributions

AP Statistics Notes Unit Two: The Normal Distributions AP Statistics Ntes Unit Tw: The Nrmal Distributins Syllabus Objectives: 1.5 The student will summarize distributins f data measuring the psitin using quartiles, percentiles, and standardized scres (z-scres).

More information

We can see from the graph above that the intersection is, i.e., [ ).

We can see from the graph above that the intersection is, i.e., [ ). MTH 111 Cllege Algebra Lecture Ntes July 2, 2014 Functin Arithmetic: With nt t much difficulty, we ntice that inputs f functins are numbers, and utputs f functins are numbers. S whatever we can d with

More information

A Scalable Recurrent Neural Network Framework for Model-free

A Scalable Recurrent Neural Network Framework for Model-free A Scalable Recurrent Neural Netwrk Framewrk fr Mdel-free POMDPs April 3, 2007 Zhenzhen Liu, Itamar Elhanany Machine Intelligence Lab Department f Electrical and Cmputer Engineering The University f Tennessee

More information

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) >

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) > Btstrap Methd > # Purpse: understand hw btstrap methd wrks > bs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(bs) > mean(bs) [1] 21.64625 > # estimate f lambda > lambda = 1/mean(bs);

More information

Section 5.8 Notes Page Exponential Growth and Decay Models; Newton s Law

Section 5.8 Notes Page Exponential Growth and Decay Models; Newton s Law Sectin 5.8 Ntes Page 1 5.8 Expnential Grwth and Decay Mdels; Newtn s Law There are many applicatins t expnential functins that we will fcus n in this sectin. First let s lk at the expnential mdel. Expnential

More information

Medium Scale Integrated (MSI) devices [Sections 2.9 and 2.10]

Medium Scale Integrated (MSI) devices [Sections 2.9 and 2.10] EECS 270, Winter 2017, Lecture 3 Page 1 f 6 Medium Scale Integrated (MSI) devices [Sectins 2.9 and 2.10] As we ve seen, it s smetimes nt reasnable t d all the design wrk at the gate-level smetimes we just

More information

Administrativia. Assignment 1 due thursday 9/23/2004 BEFORE midnight. Midterm exam 10/07/2003 in class. CS 460, Sessions 8-9 1

Administrativia. Assignment 1 due thursday 9/23/2004 BEFORE midnight. Midterm exam 10/07/2003 in class. CS 460, Sessions 8-9 1 Administrativia Assignment 1 due thursday 9/23/2004 BEFORE midnight Midterm eam 10/07/2003 in class CS 460, Sessins 8-9 1 Last time: search strategies Uninfrmed: Use nly infrmatin available in the prblem

More information

Kinetic Model Completeness

Kinetic Model Completeness 5.68J/10.652J Spring 2003 Lecture Ntes Tuesday April 15, 2003 Kinetic Mdel Cmpleteness We say a chemical kinetic mdel is cmplete fr a particular reactin cnditin when it cntains all the species and reactins

More information

Flipping Physics Lecture Notes: Simple Harmonic Motion Introduction via a Horizontal Mass-Spring System

Flipping Physics Lecture Notes: Simple Harmonic Motion Introduction via a Horizontal Mass-Spring System Flipping Physics Lecture Ntes: Simple Harmnic Mtin Intrductin via a Hrizntal Mass-Spring System A Hrizntal Mass-Spring System is where a mass is attached t a spring, riented hrizntally, and then placed

More information

Lab #3: Pendulum Period and Proportionalities

Lab #3: Pendulum Period and Proportionalities Physics 144 Chwdary Hw Things Wrk Spring 2006 Name: Partners Name(s): Intrductin Lab #3: Pendulum Perid and Prprtinalities Smetimes, it is useful t knw the dependence f ne quantity n anther, like hw the

More information

Five Whys How To Do It Better

Five Whys How To Do It Better Five Whys Definitin. As explained in the previus article, we define rt cause as simply the uncvering f hw the current prblem came int being. Fr a simple causal chain, it is the entire chain. Fr a cmplex

More information

Fall 2013 Physics 172 Recitation 3 Momentum and Springs

Fall 2013 Physics 172 Recitation 3 Momentum and Springs Fall 03 Physics 7 Recitatin 3 Mmentum and Springs Purpse: The purpse f this recitatin is t give yu experience wrking with mmentum and the mmentum update frmula. Readings: Chapter.3-.5 Learning Objectives:.3.

More information

Analysis on the Stability of Reservoir Soil Slope Based on Fuzzy Artificial Neural Network

Analysis on the Stability of Reservoir Soil Slope Based on Fuzzy Artificial Neural Network Research Jurnal f Applied Sciences, Engineering and Technlgy 5(2): 465-469, 2013 ISSN: 2040-7459; E-ISSN: 2040-7467 Maxwell Scientific Organizatin, 2013 Submitted: May 08, 2012 Accepted: May 29, 2012 Published:

More information

Checking the resolved resonance region in EXFOR database

Checking the resolved resonance region in EXFOR database Checking the reslved resnance regin in EXFOR database Gttfried Bertn Sciété de Calcul Mathématique (SCM) Oscar Cabells OECD/NEA Data Bank JEFF Meetings - Sessin JEFF Experiments Nvember 0-4, 017 Bulgne-Billancurt,

More information

Lecture 5: Equilibrium and Oscillations

Lecture 5: Equilibrium and Oscillations Lecture 5: Equilibrium and Oscillatins Energy and Mtin Last time, we fund that fr a system with energy cnserved, v = ± E U m ( ) ( ) One result we see immediately is that there is n slutin fr velcity if

More information

You need to be able to define the following terms and answer basic questions about them:

You need to be able to define the following terms and answer basic questions about them: CS440/ECE448 Sectin Q Fall 2017 Midterm Review Yu need t be able t define the fllwing terms and answer basic questins abut them: Intr t AI, agents and envirnments Pssible definitins f AI, prs and cns f

More information

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised

More information

making triangle (ie same reference angle) ). This is a standard form that will allow us all to have the X= y=

making triangle (ie same reference angle) ). This is a standard form that will allow us all to have the X= y= Intrductin t Vectrs I 21 Intrductin t Vectrs I 22 I. Determine the hrizntal and vertical cmpnents f the resultant vectr by cunting n the grid. X= y= J. Draw a mangle with hrizntal and vertical cmpnents

More information

How do scientists measure trees? What is DBH?

How do scientists measure trees? What is DBH? Hw d scientists measure trees? What is DBH? Purpse Students develp an understanding f tree size and hw scientists measure trees. Students bserve and measure tree ckies and explre the relatinship between

More information

Thermodynamics Partial Outline of Topics

Thermodynamics Partial Outline of Topics Thermdynamics Partial Outline f Tpics I. The secnd law f thermdynamics addresses the issue f spntaneity and invlves a functin called entrpy (S): If a prcess is spntaneus, then Suniverse > 0 (2 nd Law!)

More information

BLAST / HIDDEN MARKOV MODELS

BLAST / HIDDEN MARKOV MODELS CS262 (Winter 2015) Lecture 5 (January 20) Scribe: Kat Gregry BLAST / HIDDEN MARKOV MODELS BLAST CONTINUED HEURISTIC LOCAL ALIGNMENT Use Cmmnly used t search vast bilgical databases (n the rder f terabases/tetrabases)

More information

This project has received funding from the European Union s Horizon 2020 research and innovation programme under grant agreement number

This project has received funding from the European Union s Horizon 2020 research and innovation programme under grant agreement number This prject has received funding frm the Eurpean Unin s Hrizn 2020 research and innvatin prgramme under grant agreement number 727524. Credit t & http://www.h3uni.rg/ https://ec.eurpa.eu/jrc/en/publicatin/eur-scientific-andtechnical-research-reprts/behaviural-insights-appliedplicy-eurpean-reprt-2016

More information

Public Key Cryptography. Tim van der Horst & Kent Seamons

Public Key Cryptography. Tim van der Horst & Kent Seamons Public Key Cryptgraphy Tim van der Hrst & Kent Seamns Last Updated: Oct 5, 2017 Asymmetric Encryptin Why Public Key Crypt is Cl Has a linear slutin t the key distributin prblem Symmetric crypt has an expnential

More information

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA Mdelling f Clck Behaviur Dn Percival Applied Physics Labratry University f Washingtn Seattle, Washingtn, USA verheads and paper fr talk available at http://faculty.washingtn.edu/dbp/talks.html 1 Overview

More information

Electric Current and Resistance

Electric Current and Resistance Electric Current and Resistance Electric Current Electric current is the rate f flw f charge thrugh sme regin f space The SI unit f current is the ampere (A) 1 A = 1 C / s The symbl fr electric current

More information

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised

More information

AP Literature and Composition. Summer Reading Packet. Instructions and Guidelines

AP Literature and Composition. Summer Reading Packet. Instructions and Guidelines AP Literature and Cmpsitin Summer Reading Packet Instructins and Guidelines Accrding t the Cllege Bard Advanced Placement prgram: "The AP English curse in Literature and Cmpsitin shuld engage students

More information

Hypothesis Tests for One Population Mean

Hypothesis Tests for One Population Mean Hypthesis Tests fr One Ppulatin Mean Chapter 9 Ala Abdelbaki Objective Objective: T estimate the value f ne ppulatin mean Inferential statistics using statistics in rder t estimate parameters We will be

More information

Subject description processes

Subject description processes Subject representatin 6.1.2. Subject descriptin prcesses Overview Fur majr prcesses r areas f practice fr representing subjects are classificatin, subject catalging, indexing, and abstracting. The prcesses

More information

Lecture 02 CSE 40547/60547 Computing at the Nanoscale

Lecture 02 CSE 40547/60547 Computing at the Nanoscale PN Junctin Ntes: Lecture 02 CSE 40547/60547 Cmputing at the Nanscale Letʼs start with a (very) shrt review f semi-cnducting materials: - N-type material: Obtained by adding impurity with 5 valence elements

More information

Compressibility Effects

Compressibility Effects Definitin f Cmpressibility All real substances are cmpressible t sme greater r lesser extent; that is, when yu squeeze r press n them, their density will change The amunt by which a substance can be cmpressed

More information

Lab 1 The Scientific Method

Lab 1 The Scientific Method INTRODUCTION The fllwing labratry exercise is designed t give yu, the student, an pprtunity t explre unknwn systems, r universes, and hypthesize pssible rules which may gvern the behavir within them. Scientific

More information

Data Mining: Concepts and Techniques. Classification and Prediction. Chapter February 8, 2007 CSE-4412: Data Mining 1

Data Mining: Concepts and Techniques. Classification and Prediction. Chapter February 8, 2007 CSE-4412: Data Mining 1 Data Mining: Cncepts and Techniques Classificatin and Predictin Chapter 6.4-6 February 8, 2007 CSE-4412: Data Mining 1 Chapter 6 Classificatin and Predictin 1. What is classificatin? What is predictin?

More information

initially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur

initially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur Cdewrd Distributin fr Frequency Sensitive Cmpetitive Learning with One Dimensinal Input Data Aristides S. Galanpuls and Stanley C. Ahalt Department f Electrical Engineering The Ohi State University Abstract

More information

Physics 2B Chapter 23 Notes - Faraday s Law & Inductors Spring 2018

Physics 2B Chapter 23 Notes - Faraday s Law & Inductors Spring 2018 Michael Faraday lived in the Lndn area frm 1791 t 1867. He was 29 years ld when Hand Oersted, in 1820, accidentally discvered that electric current creates magnetic field. Thrugh empirical bservatin and

More information

Tree Structured Classifier

Tree Structured Classifier Tree Structured Classifier Reference: Classificatin and Regressin Trees by L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stne, Chapman & Hall, 98. A Medical Eample (CART): Predict high risk patients

More information

Name: Block: Date: Science 10: The Great Geyser Experiment A controlled experiment

Name: Block: Date: Science 10: The Great Geyser Experiment A controlled experiment Science 10: The Great Geyser Experiment A cntrlled experiment Yu will prduce a GEYSER by drpping Ments int a bttle f diet pp Sme questins t think abut are: What are yu ging t test? What are yu ging t measure?

More information

Sections 15.1 to 15.12, 16.1 and 16.2 of the textbook (Robbins-Miller) cover the materials required for this topic.

Sections 15.1 to 15.12, 16.1 and 16.2 of the textbook (Robbins-Miller) cover the materials required for this topic. Tpic : AC Fundamentals, Sinusidal Wavefrm, and Phasrs Sectins 5. t 5., 6. and 6. f the textbk (Rbbins-Miller) cver the materials required fr this tpic.. Wavefrms in electrical systems are current r vltage

More information

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS Particle Acceleratrs, 1986, Vl. 19, pp. 99-105 0031-2460/86/1904-0099/$15.00/0 1986 Grdn and Breach, Science Publishers, S.A. Printed in the United States f America ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

More information

Department of Economics, University of California, Davis Ecn 200C Micro Theory Professor Giacomo Bonanno. Insurance Markets

Department of Economics, University of California, Davis Ecn 200C Micro Theory Professor Giacomo Bonanno. Insurance Markets Department f Ecnmics, University f alifrnia, Davis Ecn 200 Micr Thery Prfessr Giacm Bnann Insurance Markets nsider an individual wh has an initial wealth f. ith sme prbability p he faces a lss f x (0

More information

Computational modeling techniques

Computational modeling techniques Cmputatinal mdeling techniques Lecture 2: Mdeling change. In Petre Department f IT, Åb Akademi http://users.ab.fi/ipetre/cmpmd/ Cntent f the lecture Basic paradigm f mdeling change Examples Linear dynamical

More information

Computational modeling techniques

Computational modeling techniques Cmputatinal mdeling techniques Lecture 4: Mdel checing fr ODE mdels In Petre Department f IT, Åb Aademi http://www.users.ab.fi/ipetre/cmpmd/ Cntent Stichimetric matrix Calculating the mass cnservatin relatins

More information

Kepler's Laws of Planetary Motion

Kepler's Laws of Planetary Motion Writing Assignment Essay n Kepler s Laws. Yu have been prvided tw shrt articles n Kepler s Three Laws f Planetary Mtin. Yu are t first read the articles t better understand what these laws are, what they

More information

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007 CS 477/677 Analysis f Algrithms Fall 2007 Dr. Gerge Bebis Curse Prject Due Date: 11/29/2007 Part1: Cmparisn f Srting Algrithms (70% f the prject grade) The bjective f the first part f the assignment is

More information

Flipping Physics Lecture Notes: Simple Harmonic Motion Introduction via a Horizontal Mass-Spring System

Flipping Physics Lecture Notes: Simple Harmonic Motion Introduction via a Horizontal Mass-Spring System Flipping Physics Lecture Ntes: Simple Harmnic Mtin Intrductin via a Hrizntal Mass-Spring System A Hrizntal Mass-Spring System is where a mass is attached t a spring, riented hrizntally, and then placed

More information

Chapter 3 Digital Transmission Fundamentals

Chapter 3 Digital Transmission Fundamentals Chapter 3 Digital Transmissin Fundamentals Errr Detectin and Crrectin CSE 3213, Winter 2010 Instructr: Frhar Frzan Mdul-2 Arithmetic Mdul 2 arithmetic is perfrmed digit y digit n inary numers. Each digit

More information

Feedforward Neural Networks

Feedforward Neural Networks Feedfrward Neural Netwrks Yagmur Gizem Cinar, Eric Gaussier AMA, LIG, Univ. Grenble Alpes 17 March 2017 Yagmur Gizem Cinar, Eric Gaussier Multilayer Perceptrns (MLP) 17 March 2017 1 / 42 Reference Bk Deep

More information

Accelerated Chemistry POGIL: Half-life

Accelerated Chemistry POGIL: Half-life Name: Date: Perid: Accelerated Chemistry POGIL: Half-life Why? Every radiistpe has a characteristic rate f decay measured by its half-life. Half-lives can be as shrt as a fractin f a secnd r as lng as

More information

NEURAL NETWORKS. Neural networks

NEURAL NETWORKS. Neural networks NEURAL NETWORKS Neural netwrks Mtivatin Humans are able t prcess cmplex tasks efficiently (perceptin, pattern recgnitin, reasning, etc.) Ability t learn frm examples Adaptability and fault tlerance Engineering

More information

Internal vs. external validity. External validity. This section is based on Stock and Watson s Chapter 9.

Internal vs. external validity. External validity. This section is based on Stock and Watson s Chapter 9. Sectin 7 Mdel Assessment This sectin is based n Stck and Watsn s Chapter 9. Internal vs. external validity Internal validity refers t whether the analysis is valid fr the ppulatin and sample being studied.

More information

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression 4th Indian Institute f Astrphysics - PennState Astrstatistics Schl July, 2013 Vainu Bappu Observatry, Kavalur Crrelatin and Regressin Rahul Ry Indian Statistical Institute, Delhi. Crrelatin Cnsider a tw

More information

Lecture 13: Markov Chain Monte Carlo. Gibbs sampling

Lecture 13: Markov Chain Monte Carlo. Gibbs sampling Lecture 13: Markv hain Mnte arl Gibbs sampling Gibbs sampling Markv chains 1 Recall: Apprximate inference using samples Main idea: we generate samples frm ur Bayes net, then cmpute prbabilities using (weighted)

More information

CHAPTER 2 Algebraic Expressions and Fundamental Operations

CHAPTER 2 Algebraic Expressions and Fundamental Operations CHAPTER Algebraic Expressins and Fundamental Operatins OBJECTIVES: 1. Algebraic Expressins. Terms. Degree. Gruping 5. Additin 6. Subtractin 7. Multiplicatin 8. Divisin Algebraic Expressin An algebraic

More information

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical mdel fr micrarray data analysis David Rssell Department f Bistatistics M.D. Andersn Cancer Center, Hustn, TX 77030, USA rsselldavid@gmail.cm

More information

Resampling Methods. Chapter 5. Chapter 5 1 / 52

Resampling Methods. Chapter 5. Chapter 5 1 / 52 Resampling Methds Chapter 5 Chapter 5 1 / 52 1 51 Validatin set apprach 2 52 Crss validatin 3 53 Btstrap Chapter 5 2 / 52 Abut Resampling An imprtant statistical tl Pretending the data as ppulatin and

More information