Computing the gradient

Computing partial derivatives

Gradient as a vector-valued function

Gradient of any function

Mnemonic for gradient

Vector of partial derivative operators

Conclusion

Gradient as an operator

Visualizing the gradient with a vector field

Vector field example

Graph interpretation

Direction of maximum gradient

Direction of steepest descent example

Direction of steepest descent

What does the length mean?

Directional derivative

Measuring a gradient

Measuring a hill

Comparing gradients

Gradient Example

Deeper gradient example

How to calculate a gradient

What does a gradient measure?

At each point, the change in z divided by the change in Y is given by the slope of this line

Let's now pick a different value for X, and keep X at this new value.

The change in z divided by the change in Y is what we refer to as the partial derivative of Z with respect to Y.

Every point on the graph has a value for the partial derivative of Z with respect to Y.

Here, green indicates a positive value, and red indicates a negative value.

Every point on the graph also has a value for the partial derivative of Z with respect to X.

Intro

Notation

Sample-based Advantage Estimation

Decomposing the L3 gradient to states and actions

MDPs with always positive rewards

Choosing a baseline

Value function estimation

Fitting a baseline

Experimental Results

Summary

Introduction

Recap

Using training data

Cost functions

Gradient descent

More on gradient vectors

Gradient descent recap

Analyzing the network

Learning more

Lisha Li interview

Closing thoughts

Awesome song and introduction

Main ideas behind Gradient Descent

Gradient Descent optimization of a single variable, part 1

An important note about why we use Gradient Descent

Gradient Descent optimization of a single variable, part 2

Review of concepts covered so far

Gradient Descent optimization of two (or more) variables

A note about Loss Functions

Gradient Descent algorithm

Stochastic Gradient Descent

Intro

What are gradients?

Using gradients in Figma

Linear gradients

Radial gradients

Angular gradients

Diamond gradients

The four gradient types

Figma color styles

Computing the gradient

Directional derivative intuition

Evaluating the gradient at a point

What the dot product represents

How to interpret the dot product

When is the dot product maximized?

The direction of steepest descent is the gradient itself

Interpreting gradient length

Directional derivative

Directional derivative in direction of gradient

Intro

Y-intercept

Gradient formula

Rise and Run

Finding the gradient of a line

Gradient as a measure of steepness

Negative gradient

Positive gradient

Finding the Gradient

Conclusion

Enabling the gradient options

Gradient Editor

Selecting gradient colours

Built in presets

How to remove banding

Other gradient properties

Gradient styles

Overlaying a gradient onto a graphic

Blending modes

Adjusting weights and biases

Learning rate

Gradient Descent Example

Training new data

Three types of Gradient Descent

Stochastic Gradient Descent

Intro

Naming the gradient

Conclusion