Other articles


  1. Analysis of variance

    Analysis of Variance

    The mathematics of the last century worth of experiment design.

    Probably the least sexy thing in statistics and as such, usually taught by the least interesting professor in the department, or at least one who couldn't find an interesting enough excuse to get out of it ...

    read more
  2. Privacy

    Privacy

    A frequent conversation I'm having with people, in the post Snowden era is "oh no, my privacy is gone, there's nothing I can do to keep it."

    I think this is an uneccessarily unproductive way of thinking about it; sure maintaining a paranoid cold-war-bunker-style absolute secrecy is ...

    read more
  3. Manifold learning

    As in - handling your high-dimensional data by trying to discover a low-dimensional manifold that contains it. There are a million different versions of this. See also "Functional regression", where the manifold isn'et ever necessarily low dimensional,

    To understand: How does this relate to information geometry? (Related, but ...

    read more
  4. pseudorandomness

    Pseudorandomness

    Not just PRNGs, not Erdős-style probabilistic methods, though those are both closely related. Rather, the intersection between probability, ignorance, and algorithms, butting up against computability theory. When is the relation between things sufficiently unstructured that we may treat it as random? Stochastic approximations to deterministic algorithsm. Kolmogorov complexity. Compressibility ...

    read more
  5. Branching Processes

    A class of stochastic models generalising the Galton Watson process that I am mildly obsessed with.

    • Caballero, M.-E., Lambert, A., & Bravo, G. U. (2009). Proof(s) of the Lamperti representation of Continuous-State Branching Processes. Probability Surveys, 6, 62–89. DOI. Online.
    • Crisan, D., Del Moral, P., & Lyons ...
    read more
  6. Kernel methods

    In the sense of the "Kernel trick". Hilbert spaces and product kernels and reproducing kernel hilbert spaces. Not desnity estimation kernels (related), nor the dozens of other definitions.

    A neat trick to upgrade your old boring linear algebra on finite dimenisonal spaces to sexy new curvy algebra on ...

    read more
  7. Expectation maximisation

    Expectation maximisation

    Something else I should learn; An iterative method for finding local solutions to a common class of otherwise intractable ML estiamtion problems, especially missing data type problems. Apprently it avoid derivatives and therefore that awful moment when you are implementing some optimisation or other and the computer blithely ...

    read more
  8. iterated function systems

    Iterated function systems

    In the display case between l-systems and classic fractals like the Mandelbrot set, we have IFSs. There is a lot of definitional overlap between these things; mostly the difference is about where they are used and the algorithms used to draw them.

    • Why does no-one (seem to ...
    read more

Page 1 / 10 »

social