The Baron's most recent wager with Sir R----- set him the challenge of being the last to remove a horizontally, vertically or diagonally adjacent pair of draughts from a five by five square of them, with the Baron first taking a single draught and Sir R----- and he thereafter taking turns to remove such pairs.

When I heard these rules I was reminded of the game of Cram and could see that, just like it, the key to figuring the outcome is to recognise that the Baron could always have kept the remaining draughts in a state of symmetry, thereby ensuring that however Sir R----- had chosen he shall subsequently have been free to make a symmetrically opposing choice.

Over the last few months we have seen how we can use a sequence of Householder transformations followed by a sequence of shifted Givens rotations to efficiently find the spectral decomposition of a symmetric real matrix M, formed from a matrix V and a diagonal matrix Λ satisfying

    M × V = V × Λ

implying that the columns of V are the unit eigenvectors of M and their associated elements on the diagonal of Λ are their eigenvalues so that

    V × V^T = I

where I is the identity matrix, and therefore

    M = V × Λ × V^T

From a mathematical perspective the combination of Householder transformations and shifted Givens rotations is particularly appealing, converging on the spectral decomposition after relatively few matrix multiplications, but from an implementation perspective using ak.matrix multiplication operations is less than satisfactory since it wastefully creates new ak.matrix objects at each step and so in this post we shall start to see how we can do better.

Over the course of the year, my fellow students and I have been experimenting with an ethereal orrery which models the motion of heavenly bodies using nought but Sir N-----'s laws of gravitation and motion. Whilst the consequences of those laws are not generally subject to solution by mathematical reckoning, we were able to approximate them with a scheme that admitted errors of the order of the sixth power of the steps in time by which we advanced the positions of those bodies.
We have thus far employed it to model the solar system itself, uniformly distributed bodies of matter and the accretion of bodies that are close to Earth's orbit about the Sun. Whilst we were most satisfied by its behaviour, I should now like to report upon an altogether more surprising consequence of its engine's action.

Last time we saw how we could use a sequence of Householder transformations to reduce a symmetric real matrix M to a symmetric tridiagonal matrix, having zeros everywhere other than upon the leading, upper and lower diagonals, which we could then further reduce to a diagonal matrix Λ using a sequence of Givens rotations to iteratively transform the elements upon the upper and lower diagonals to zero so that the columns of the accumulated transformations V were the unit eigenvectors of M and the elements on the leading diagonal of the result were their associated eigenvalues, satisfying

    M × V = V × Λ

and, since the transpose of V is its own inverse

    M = V × Λ × V^T

which is known as the spectral decomposition of M.
Unfortunately, the way that we used Givens rotations to diagonalise tridiagonal symmetric matrices wasn't particularly efficient and I concluded by stating that it could be significantly improved with a relatively minor change. In this post we shall see what it is and why it works.

Hello there Sir R-----! Come join me by the hearth for a dram of warming spirits! I trust that this cold spell has not chilled your desire for a wager?

Good man! Good man!

I must say that the contrast between the warmth of this fire and the frost outside brings most vividly to my mind an occasion during my tenure as the Empress's ambassador to the land of Oz; specifically the time that I attended King Quadling Rex's winter masked ball during which his southern palace was overrun by an infestation of Snobbles!

Some years ago we saw how we could use the Jacobi algorithm to find the eigensystem of a real valued symmetric matrix M, which is defined as the set of pairs of non-zero vectors v_i and scalars λ_i that satisfy

    M × v_i = λ_i × v_i

known as the eigenvectors and the eigenvalues respectively, with the vectors typically restricted to those of unit length in which case we can define its spectral decomposition as the product

    M = V × Λ × V^T

where the columns of V are the unit eigenvectors, Λ is a diagonal matrix whose i^th diagonal element is the eigenvalue associated with the i^th column of V and the T superscript denotes the transpose, in which the rows and columns of the matrix are swapped.
You may recall that this is a particularly convenient representation of the matrix since we can use it to generalise any scalar function to it with

    f(M) = V × f(Λ) × V^T

where f(Λ) is the diagonal matrix whose i^th diagonal element is the result of applying f to the i^th diagonal element of Λ.
You may also recall that I suggested that there's a more efficient way to find eigensystems and I think that it's high time that we took a look at it.

The latest wager that the Baron put to Sir R----- had them competing to first chalk a triangle between three of eight coins, with Sir R----- having the prize if neither of them managed to do so. I immediately recognised this as the game known as Clique and consequently that Sir R-----'s chances could be reckoned by applying the pigeonhole principle and the tactic of strategy stealing. Indeed, I said as much to the Baron but I got the distinct impression that he wasn't really listening.

For several months we've for been taking a look at cluster analysis which seeks to partition sets of data into subsets of similar data, known as clusters. Most recently we have focused our attention on hierarchical clusterings, which are sequences of sets of clusters in which pairs of data that belong to the same cluster at one step belong to the same cluster in the next step.
A simple way of constructing them is to initially place each datum in its own cluster and then iteratively merge the closest pairs of clusters in each clustering to produce the next one in the sequence, stopping when all of the data belong to a single cluster. We have considered three ways of measuring the distance between pairs of clusters, the average distance between their members, the distance between their closest members and the distance between their farthest members, known as average linkage, single linkage and complete linkage respectively, and implemented a reasonably efficient algorithm for generating hierarchical clusterings defined with them, using a min-heap structure to cache the distances between clusters.
Finally, I claimed that there is a more efficient algorithm for generating single linkage hierarchical clusterings that would make the sorting of clusters by size in our ak.clustering type too expensive and so last time we implemented the ak.rawClustering type to represent clusterings without sorting their clusters which we shall now use in the implementation of that algorithm.

Recently, my fellow students and I constructed a mathematical orrery which modelled the motion of heavenly bodies employing Sir N-----'s laws of gravitation and motion, rather than clockwork, as its engine. Those laws state that bodies are attracted toward each other with a force proportional to the product of their masses divided by the square of the distance between them, that a body will remain at rest or in constant motion unless a force acts upon it, that if a force acts upon it then it will be accelerated in the direction of that force at a rate proportional to its strength divided by its mass and that, if so, it will reciprocate with an opposing force of equal strength.
Its operation was most satisfactory, which set us to wondering whether we might use its engine to investigate the motions of entirely hypothetical arrangements of heavenly bodies and I should now like to report upon our progress in doing so.

Last month we saw how we could efficiently generate hierarchical clusterings, which are sequences of sets of clusters, which are themselves subsets of a set of data that each contain elements that are similar to each other, such that if a pair of data are in the same clustering at one step then they must be in the same clustering in the next which will always be the case if we move from one step to the next by merging the closest pairs of clusters. Specifically, we used our ak.minHeap implementation of the min-heap structure to cache the distances between clusters, saving us the expense of recalculating them for clusters that don't change from one step in the hierarchy to the next.
Recall that we used three different schemes for calculating the distance between a pair of clusters, the average distance between their members, known as average linkage, the distance between their closest members, known as single linkage, and the distance between their farthest members, known as complete linkage, and that I concluded by noting that our algorithm was about as efficient as possible in general but that there is a much more efficient scheme for single linkage clusterings; efficient enough that sorting the clusters in each clustering by size would be the most costly operation and so in this post we shall implement objects to represent clusterings that don't do that.