Gradient Descent Step
MediumMachine LearningMath
Description
Given a weight vector, a gradient vector of the same length, and a learning rate, return the updated weights after one gradient-descent step: each weight minus the learning rate times its gradient. Round each result to 4 decimal places.
Examples
Input:
[1,2,3], [0.1,0.2,0.3], 1Output:
[0.9,1.8,2.7]Explanation:
Each weight moves against its gradient by a step proportional to the learning rate, nudging the model toward lower loss.
Input:
[5,5], [1,1], 0.5Output:
[4.5,4.5]Explanation:
Each weight moves against its gradient by a step proportional to the learning rate, nudging the model toward lower loss.
Input:
[0,0], [2,-2], 0.1Output:
[-0.2,0.2]Explanation:
Each weight moves against its gradient by a step proportional to the learning rate, nudging the model toward lower loss.
Constraints
- •
1 ≤ length ≤ 10⁴ - •
learning rate > 0