Machine Learning

Introduction to Machine Learning

Challenge Question

https://dl.dropbox.com/s/w1r1odresud2h3j/Screenshot%202017-03-06%2020.49.23.png
  • Answer C

  • Decision Trees are understandable and easy to explain.

k-Nearest Neighbors

https://dl.dropbox.com/s/ud6yqjfj82gwaak/Screenshot%202017-03-06%2020.52.30.png https://dl.dropbox.com/s/l9id3i0cdplwcci/Screenshot%202017-03-06%2020.53.37.png
  • Normal Bite (O)

  • Cookie Cutter Bite (+)

  • Find the nearest example in the training data-set and apply that label.

Cross Validation

Cross Validation

Note: When applying k-fold cross-validation, you usually need to select the training/test samples on each iteration randomly. But in certain cases, such as for time-series or sequence data, random selection not be a valid approach.

https://dl.dropbox.com/s/nnc4tblx32cr0km/Screenshot%202017-03-06%2020.57.54.png https://dl.dropbox.com/s/4nkz70iz59ymo1n/Screenshot%202017-03-06%2021.00.00.png

Cross Validation Quiz

https://dl.dropbox.com/s/sjs3x7voadlei7b/Screenshot%202017-03-06%2021.06.56.png
  • It is claimed that the number of pizza deliveries to the Pentagon was used to predict the start of the first Iraq War.

The Gaussian Distribution

https://dl.dropbox.com/s/78n5znh0olbmt2s/Screenshot%202017-03-06%2021.21.29.png

Central Limit Theorem

Grasshoppers Vs Katydids

https://dl.dropbox.com/s/tx8xmx3nx3op7dq/Screenshot%202017-03-06%2022.36.22.png https://dl.dropbox.com/s/gb4cj5yfiudg4ol/Screenshot%202017-03-06%2022.36.58.png?dl=0

Gaussians Quiz

https://dl.dropbox.com/s/dhqm7z77ofolisy/Screenshot%202017-03-06%2022.40.13.png?dl=0

Decision Boundaries

https://dl.dropbox.com/s/00lpr8fmmggc5hc/Screenshot%202017-03-06%2023.20.29.png

Recognition Quiz

https://dl.dropbox.com/s/47i4wtdto8xmuoy/Screenshot%202017-03-06%2023.22.44.png?dl=0

Decision Boundaries

https://dl.dropbox.com/s/tc6iio8ths44oig/Screenshot%202017-03-06%2023.24.27.png?dl=0

Error

https://dl.dropbox.com/s/ks4pkqepgjer9np/Screenshot%202017-03-06%2023.24.50.png?dl=0 https://dl.dropbox.com/s/c1vxqsfpt9wd43t/Screenshot%202017-03-06%2023.30.54.png?dl=0 https://dl.dropbox.com/s/biwkj3umy1hayyo/Screenshot%202017-03-06%2023.31.29.png?dl=0

Bayes Classifier

https://dl.dropbox.com/s/evcb58l7tmccy3i/Screenshot%202017-03-06%2023.33.50.png?dl=0

Bayes Rule Quiz

https://dl.dropbox.com/s/850lpklbblscalb/Screenshot%202017-03-06%2023.35.19.png?dl=0

Naive Bayes

https://dl.dropbox.com/s/bqmeuxls073lh9a/Screenshot%202017-03-06%2023.36.27.png?dl=0 https://dl.dropbox.com/s/8i3bvh308pe81kk/Screenshot%202017-03-06%2023.37.20.png?dl=0

Maximum Likelihood

https://dl.dropbox.com/s/25nz51aehsprnqf/Screenshot%202017-03-06%2023.38.36.png?dl=0

Naive Bayes Quiz

https://dl.dropbox.com/s/hraobnsfp5zp030/Screenshot%202017-03-06%2023.50.26.png https://dl.dropbox.com/s/puu6p1k6fuqynw2/Screenshot%202017-03-06%2023.51.41.png?dl=0

No Free Lunch

Naive Bayes vs kNN

https://dl.dropbox.com/s/ixtte5v07v3yrpf/Screenshot%202017-03-06%2023.53.25.png

Using a Mixture of Gaussians

https://dl.dropbox.com/s/edwjyihfutln7q8/Screenshot%202017-03-06%2023.54.41.png?dl=0
  • Kernel Density Estimation.

  • Cross-Validation to avoid Overfitting.

Generalizations

https://dl.dropbox.com/s/pdheypcaalpwf3a/Screenshot%202017-03-06%2023.56.17.png?dl=0

Decision Tree with Discrete Information

https://dl.dropbox.com/s/ykm48yq0jkcuewo/Screenshot%202017-03-06%2023.58.57.png?dl=0

Decision Tree Quiz 1

https://dl.dropbox.com/s/obhruvfqd91lcvm/Screenshot%202017-03-07%2000.00.11.png?dl=0

Decision Trees with Continuos Information

https://dl.dropbox.com/s/c9a83ynb40qxd3g/Screenshot%202017-03-07%2000.01.24.png?dl=0

Minimum Description Length

https://dl.dropbox.com/s/4q856rzay7r0ngv/Screenshot%202017-03-07%2000.03.08.png?dl=0 https://dl.dropbox.com/s/526gyr8zylj25rv/Screenshot%202017-03-07%2000.03.59.png https://dl.dropbox.com/s/sd9kqoxiimljn4l/Screenshot%202017-03-07%2007.57.24.png https://dl.dropbox.com/s/42olbo256dfgur5/Screenshot%202017-03-07%2007.58.19.png?dl=0 https://dl.dropbox.com/s/m9k9cybt00nvz31/Screenshot%202017-03-07%2007.58.55.png?dl=0

Entropy

https://dl.dropbox.com/s/ujok961tist0sni/Screenshot%202017-03-07%2007.59.18.png?dl=0 https://dl.dropbox.com/s/i7wl7s4jyc35wbp/Screenshot%202017-03-10%2011.32.28.png
  • We will use entropy to determine the decision tree branching.

https://dl.dropbox.com/s/y044t2rm288rozm/Screenshot%202017-03-07%2008.00.10.png?dl=0

Information Gain

https://dl.dropbox.com/s/fu1uae0gs0tb7f5/Screenshot%202017-03-07%2008.00.39.png?dl=0
  • then we can figure out the most important attributes.

https://dl.dropbox.com/s/magt35rncixw2sd/Screenshot%202017-03-07%2008.01.47.png?dl=0 https://dl.dropbox.com/s/om0jd14225e9vz5/Screenshot%202017-03-07%2008.03.18.png?dl=0
  • We cna use the same attribute at multiple levels in the decisions trees.

Decision Tree Quiz 2

https://dl.dropbox.com/s/9sk8x6zwb7nod0z/Screenshot%202017-03-07%2008.04.41.png?dl=0 https://dl.dropbox.com/s/e8rq9pvogkibarh/Screenshot%202017-03-07%2008.04.12.png?dl=0

Random Forests

https://dl.dropbox.com/s/6z0zae804rt8837/Screenshot%202017-03-07%2008.06.52.png?dl=0 https://dl.dropbox.com/s/cuk7bkapxkl5pd8/Screenshot%202017-03-07%2008.07.08.png?dl=0

Boosting

https://dl.dropbox.com/s/wef0e73ksiw51yu/Screenshot%202017-03-07%2008.11.37.png?dl=0 https://dl.dropbox.com/s/icemif3z3r4l9xb/Screenshot%202017-03-07%2008.11.59.png?dl=0 https://dl.dropbox.com/s/en88x3q4y9hip1i/Screenshot%202017-03-07%2008.12.20.png?dl=0 https://dl.dropbox.com/s/rw7tqle7fiwophi/Screenshot%202017-03-07%2008.12.39.png?dl=0 https://dl.dropbox.com/s/lh9kiv897rl23uj/Screenshot%202017-03-07%2008.13.15.png?dl=0

Boosting Quiz

https://dl.dropbox.com/s/azjtyoz7u2lyikm/Screenshot%202017-03-07%2008.14.55.png?dl=0 https://dl.dropbox.com/s/hlc0hujt8zhtvyi/Screenshot%202017-03-07%2008.15.39.png?dl=0

Neural Nets

https://dl.dropbox.com/s/yuqq33zkq6vlai1/Screenshot%202017-03-07%2008.16.44.png?dl=0 https://dl.dropbox.com/s/z7fl7xy3audmtzi/Screenshot%202017-03-07%2008.17.18.png?dl=0 https://dl.dropbox.com/s/fe714pyoswn2ju6/Screenshot%202017-03-07%2008.17.33.png?dl=0 https://dl.dropbox.com/s/c55a9kx9gr9f761/Screenshot%202017-03-07%2008.17.54.png?dl=0

Neural Nets Quiz

https://dl.dropbox.com/s/i1whn2twi04oyvf/Screenshot%202017-03-07%2008.18.22.png?dl=0
Quiz: Neural Nets Quiz
Fill in the truth table for NOR and find weights such that:

a = { true if w0 + i1 w1 + i2 w2 > 0, else false }

Truth table

Enter 1 for True, and 0 (or leave blank) for False in each cell.
All combinations of i1 and i2 must be specified.

Weights

Each weight must be a number between 0.0 and 1.0, accurate to one or two decimal places.
w1 and w2 are the input weights corresponding to i1 and i2 respectively.
w0 is the bias weight.

Activation function
Choose the simplest activation function that can be used to capture this relationship.
https://dl.dropbox.com/s/kwmjhucwyt96500/Screenshot%202017-03-07%2008.19.21.png?dl=0

Multilayer Nets

https://dl.dropbox.com/s/sp3cnnpmg4uh80c/Screenshot%202017-03-07%2008.20.01.png?dl=0

Perceptron Learning

https://dl.dropbox.com/s/m1wktevt5o9vyy0/Screenshot%202017-03-07%2008.21.42.png?dl=0 https://dl.dropbox.com/s/q4dn1z97pnoc5e2/Screenshot%202017-03-07%2008.22.20.png?dl=0 https://dl.dropbox.com/s/iu6wbzpw9f0m33y/Screenshot%202017-03-07%2008.24.17.png?dl=0

Expressiveness of Preceptron

https://dl.dropbox.com/s/6g6k85b3lskqt14/Screenshot%202017-03-07%2008.25.44.png?dl=0 https://dl.dropbox.com/s/1grs2slb8t0on5f/Screenshot%202017-03-07%2008.25.57.png?dl=0

Multilayer Perceptron

https://dl.dropbox.com/s/tnfpepvmprbj2hq/Screenshot%202017-03-07%2008.26.33.png?dl=0 https://dl.dropbox.com/s/8lz2epthpuskypu/Screenshot%202017-03-07%2008.26.55.png?dl=0

Back-Propagation

https://dl.dropbox.com/s/ax8ohieruif9c0o/Screenshot%202017-03-07%2008.27.31.png?dl=0 https://dl.dropbox.com/s/xfz47n55x9u7osv/Screenshot%202017-03-07%2008.29.11.png?dl=0 https://dl.dropbox.com/s/ox29qofh5ld0tnb/Screenshot%202017-03-07%2008.29.27.png?dl=0

Deep Learning

Unsupervised Learning

k-Means and EM

https://dl.dropbox.com/s/pzw3ebew6gty3pi/Screenshot%202017-03-07%2008.32.50.png?dl=0 https://dl.dropbox.com/s/nva08m719r9a96e/Screenshot%202017-03-07%2008.33.10.png?dl=0 https://dl.dropbox.com/s/l53k69rjcu2ua5c/Screenshot%202017-03-07%2008.33.32.png?dl=0 https://dl.dropbox.com/s/w3l1dshvy6gr9qx/Screenshot%202017-03-07%2008.33.54.png?dl=0

EM and Mixture of Gaussians

https://dl.dropbox.com/s/xccinv5ayqlzu64/Screenshot%202017-03-07%2008.35.21.png?dl=0 https://dl.dropbox.com/s/ystxkm0uqfv4rav/Screenshot%202017-03-07%2008.35.44.png?dl=0

Resources