alleryStart learning →The gradient-descent update rule for parameter θⱼ in linear regression is
Aθⱼ := θⱼ + α · ∂J/∂θⱼ
Bθⱼ := θⱼ − α · ∂J/∂θⱼ
Cθⱼ := α · θⱼ
Dθⱼ := θⱼ × J(θ)
Answer & Solution
Correct answer: B. θⱼ := θⱼ − α · ∂J/∂θⱼ
Gradient descent steps DOWN the gradient: θⱼ := θⱼ − α · ∂J/∂θⱼ.
The minus sign is what makes it descent rather than ascent. α is the learning rate.
Keep practising Linear Regression and Supervised Learning
Browse all Linear Regression and Supervised Learning questions for Claude, or explore more Aifoundations topics.
Related questions
Split batch from stochastic descentThe resulting per-example log-likelihood is, up to constantsClosed-form normal equations require XᵀX to beLogistic regression uses which hypothesis function?The PROBABILISTIC justification for the least-squares cost assumes the noise ε⁽ⁱ⁾ in y⁽ⁱ⁾ The design matrix X for n training examples with d features (plus intercept) has shapeThe closed-form 'normal equations' solution for least-squares linear regression isThe LMS (Widrow–Hoff) update rule on a single training example is