Understanding Coordinate Descent
Course link: https://www.coursera.org/learn/ml-regression
let's just have a little aside on the coordinate decent algorithm, and then we're gonna describe how to apply coordinate descent to solving our lasso objective. So, our goal here is to minimize sub function g. So, this is the same objective that we have whether we are talking about our closed form solution, gradient descent, or this coordinate descent algorithm. But, let me just be very explicit, where. We're saying we wanna minimize over all possible w some g(w), where here, we're assuming g(w) is function of multiple variables. Let's call it g(w0,w1,...,wD). So this W we are trying to write in some bold font here. And often, minimizing over a large set of variables can be a very challenging problem. But in contrast, often it's possible to think about optimizing just a single dimension, keeping all of the other dimensions fixed. So easy for each coordinate when keeping others fixed, because that turns into just a 1D optimization problem. And so, that's the motivation behind coordinate decent, where the coordinate descent algorithm, it's really intuitive.
Видео Understanding Coordinate Descent канала Machine Learning TV
let's just have a little aside on the coordinate decent algorithm, and then we're gonna describe how to apply coordinate descent to solving our lasso objective. So, our goal here is to minimize sub function g. So, this is the same objective that we have whether we are talking about our closed form solution, gradient descent, or this coordinate descent algorithm. But, let me just be very explicit, where. We're saying we wanna minimize over all possible w some g(w), where here, we're assuming g(w) is function of multiple variables. Let's call it g(w0,w1,...,wD). So this W we are trying to write in some bold font here. And often, minimizing over a large set of variables can be a very challenging problem. But in contrast, often it's possible to think about optimizing just a single dimension, keeping all of the other dimensions fixed. So easy for each coordinate when keeping others fixed, because that turns into just a 1D optimization problem. And so, that's the motivation behind coordinate decent, where the coordinate descent algorithm, it's really intuitive.
Видео Understanding Coordinate Descent канала Machine Learning TV
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![Large Scale Datasets and Very Deep Neural Networks - Deep Learning with Python](https://i.ytimg.com/vi/09Q4pXQUk1c/default.jpg)
![Linear Regression by Tom Mitchell](https://i.ytimg.com/vi/pQwF_qUYYR8/default.jpg)
![Loading Pre trained Models with Theano - Deep Learning with Python](https://i.ytimg.com/vi/YkGieAgSWho/default.jpg)
![Semantics & Factorization - Stanford University](https://i.ytimg.com/vi/1RgkCkN6IM0/default.jpg)
![Deep Features - University of Washington](https://i.ytimg.com/vi/HVbUD9aA_Ys/default.jpg)
![Challenges of Deep learning - University of Washington](https://i.ytimg.com/vi/WWpj5Qs9r18/default.jpg)
![What Never Ending Learning (NELL) Really is? - Tom Mitchell](https://i.ytimg.com/vi/MUMkrhrDmqQ/default.jpg)
![Making Sense of the World with Deep Learning - Facebook](https://i.ytimg.com/vi/M8qdcOxDxgA/default.jpg)
![Understanding ChatGPT and LLMs from Scratch - Part 2](https://i.ytimg.com/vi/FsDVOYY3kD8/default.jpg)
![Basic Neural Nets with TensorFlow](https://i.ytimg.com/vi/rzilPq7xXC4/default.jpg)
![Understanding Deep Learning with Theano](https://i.ytimg.com/vi/K6m-iOU3iqk/default.jpg)
![Convolutional Weights as Image Features ( Deep Learning with TensorFlow)](https://i.ytimg.com/vi/rrOgPiqYu6s/default.jpg)
![Keras Behind the Scenes - Deep Learning with Python](https://i.ytimg.com/vi/Jy1Fkdg6npY/default.jpg)
![Python Libraries for Machine Learning You Must Know!](https://i.ytimg.com/vi/TyKzBoEaeEM/default.jpg)
![Naive Bayes by Tom Mitchell](https://i.ytimg.com/vi/qZHTxW_25rc/default.jpg)
![Lecture 18 320x240](https://i.ytimg.com/vi/rkbW3OpWP50/default.jpg)
![Examples of Deep Learning in Computer Vision - University of Washington](https://i.ytimg.com/vi/wCVf6bwG8gc/default.jpg)
![Probability and Estimation by Tom Mitchell](https://i.ytimg.com/vi/XEEOV9Solpo/default.jpg)
![Towards Practical Machine Learning with Differential Privacy and Beyond](https://i.ytimg.com/vi/THmTZ7WOkbY/default.jpg)
![Template Models: Plate Models - Stanford University](https://i.ytimg.com/vi/fRhaGgMjgso/default.jpg)
![Page Ranking: Web as a Graph (Stanford University 2019)](https://i.ytimg.com/vi/-zq9-6RbKZc/default.jpg)