Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

My question is at the bottom. (Most of the descriptive words come from Chris. Bishop's Neural Networks for Pattern Recognition)

Let $w$ be the weight vector of the neural network and $E$ the error function.

According to the Robbins-Monro algorithm, this sequence: $$w_{kj}^{(r+1)}=w_{kj}^{(r)}-\eta\left.\frac{\partial E}{\partial w_{kj}}\right|_{w^{(r)}}$$ will converge to a limit, for which: $$\frac{\partial E}{\partial w_{kj}}=0.$$

In general the error function is given by a sum of terms, each of which is calculated using one of the patterns from the training set, so that $$E=\sum_nE^n(w)$$ And in applications we just update the weight vector using one pattern at a time $$w_{kj}^{(r+1)}=w_{kj}^{(r)}-\eta\frac{\partial E^n}{\partial w_{kj}}$$

My question is: Why will the algorithm converge using the last formula? Once we use it to update the $w$, the value of $w$ is changed, and I can't prove the convergence using $$\frac{\partial E}{\partial w_{kj}}=\sum_n \frac{\partial E^n}{\partial w_{kj}}$$

share|cite|improve this question
I don't know anything about neural networks, but this looks a multidimensional Newton's method, see . Two common strategies to prove convergence of methods like this are 1) show the iteration is a contraction mapping, or 2) Find some quantity that decreases by at least a fixed fraction every iteration that is zero at the true solution. It will be difficult to help much more without knowing more properties of E. – Nick Alger Apr 8 '12 at 4:20

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.