Machine Learning Lecture 15 '(Linear) Support Vector Machines continued' -Cornell CS4780 SP17

Показать описание

Lecture Notes:

Рекомендации по теме

Комментарии

Not sure why this playlist doesn't come on top when one searches for ML on youtube. Andrew Ng might be a good researcher but he's not a very good teacher. Killian teaches in such a good manner, that I never once felt bored of felt as if I'm studying. Thanks Killian, You're a Gem.

saransh

I am glad that I choose your course to learn the theory of ML. May God bless you and your family! Thank you!

Anuarlogon

Thanks a lot for making this high quality content publicly available - much appreciated!

matthieuglotz

Your natural intelligence is trained to check for mic before every class after failure in one :):)

kirtanpatel

cool man, you shows by graph how excellent the SVM is :D:D:D very clear and vivid graph.

gregmakov

Thanks for the awesome lectures! We have a weekly "book club" of Cornell MAE alumn teaching ourselves ML based on your course. We've all been wondering where you get your sweaters?

lawrencelenkin

I want to know about svm hinge loss function and svm loss function more.. does anyone know a good recourse?

ehfo

Hi Prof
I have 2 doubts:
1- Does SVM outputs only 1 hyperplane? I guess some planes passing through midpoint of line joining support vectors also satisfies all equations.
2- In your lecture notes on changing the value of "c" why only magnitude of "w" is changing and no change in the direction of "w" is observed?

ayushmalik

Why are we considering wTw as the regularizer in the hinge function when originally wTw was being minimized (without considering any slack).
Shouldn't the introduction of slack be the regulariser here

uddishnegi

Is margin defined differently in the case of svm with soft constraints, i.e. whether the points which don't satisfy y_i (transpose(w) * x_i +b) > = 1 are considered while calculating margin (which is min d (where d is distance of each point to a given hyperplane)) ?

sudhanshuvashisht

why does W get smaller as C decreases?

ehfo

logloss is the (negative) loglikelihood of P(Y|X) (i think). Do the other losses correspond to some modeling of the P(Y|X) as well?

deltasun

Machine Learning Lecture 15 '(Linear) Support Vector Machines continued' -Cornell CS4780 SP17

Machine Learning Lecture 15 '(Linear) Support Vector Machines continued' -Cornell CS4780 S...

Machine Learning Lecture 14 '(Linear) Support Vector Machines' -Cornell CS4780 SP17

Lecture 15 : Linear Regression

Linear models and simulations | MIT Computational Thinking Spring 2021 | Lecture 15

Pattern Recognition [PR] Episode 15 - Linear Discriminant Analysis - Examples

15 Linear Regression

Eigen values and Eigen vectors (PCA): Dimensionality reduction Lecture 15@ Applied AI Course

Machine Learning Foundations Course – Regression Analysis

Electrical Machines - Transformers Exam Insights Series by Chetan Sir | Next Engineer #sscje #rrbje

15. Linear Programming: LP, reductions, Simplex

ML Lecture 15: Unsupervised Learning - Neighbor Embedding

Machine Learning Lecture 26 'Gaussian Processes' -Cornell CS4780 SP17

Deep Learning Lecture 15: Deep Reinforcement Learning - Policy search

Linear Regression Algorithm – Solved Numerical Example in Machine Learning by Mahesh Huddar

Lecture 15 | Efficient Methods and Hardware for Deep Learning

Machine Learning Course for Beginners

CS231n Winter 2016: Lecture 15: Invited Talk by Jeff Dean

Cornell CS 5787: Applied Machine Learning. Lecture 5b. Part 1: Probabilistic Linear Regression

undergraduate machine learning 17: Linear prediction

SP15 Lecture 21 Part 2 Linear Classifier

Lecture 15 Time Series Modeling

Linear Regression Solved Numerical Part-1 Explained in Hindi l Machine Learning Course

Linear Regression Explained in Hindi ll Machine Learning Course

Lecture 15 | Applied Linear Algebra | Vector Properties | Prof AK Jagannatham