Skip to content

Building Babylon

Scribbles on maths and machine learning

Posted on 17 February 20158 May 2023 by Benjamin

Maximum Likelihood Estimation for Non-negative Matrix Factorisation and the generalised Kullback-Leibler divergence

In response to my own question at Cross Validated.

CategoriesMath, Matrix factorisation

Post navigation

Next PostNext An overview of word2vec

Recent Posts

  • Deep Differentiable Logic Gate Networks & their (potential) relevance to ZKML
  • Elliptic curve pairings: a first example
  • A toy elliptic curve over a finite field
  • Siegelmann & Sontag’s “On the Computational Power of Neural Nets”
  • Graph embeddings in Hyperbolic Space

Pages

  • Publications
  • Representation theory
  • About
  • Posts
  • Machine Learning Seminar

Recent comments

  1. Don’t interpret linear hidden units, they don’t exist. – Building Babylon on Orthogonal transformations and gradient updates
  2. Squared pair-wise differences estimate the variance – Building Babylon on Estimation Theory and the Bias-Variance Trade-Off
  3. Matrix Factorisation and the Eigendecomposition of the Gram Matrix – Building Babylon on Eigendecomposition of real, symmetric matrices
  4. Marginal and Conditional Distributions of the Multivariate Gaussian – Building Babylon on Block Multiplication of Matrices
  5. Don’t interpret linear hidden units, they don’t exist. – Building Babylon on Descartes’ headache: rotation of space is transpose to rotation of co-ordinate system

Categories

  • Datasets
  • Elliptic curves
  • Geometry
  • gradient descent
  • History
  • Language modeling
  • Math
  • Matrix factorisation
  • Optimisation
  • Pontifications
  • Recommendation
  • Reviews
  • Talks
  • Teaching
  • ZKML
Proudly powered by WordPress