January 2020 archive

Jan 30 2020

Book review: How the States got their Shapes by Mark Stein

how_the_statesHow the States got their Shapes by Mark Stein is that book that does exactly what it says on the cover: explain the origin of the shapes of the states of the United States. The book starts with some broad brush strokes that underpin the shaping of many states before going through each State in alphabetical order.

States are not strictly comparable with European nations but it is interesting to compare the never-straight borders of Europe with the regularity of particularly Western states. To a British European the events described in the book are all terribly recent – much of the action occurs during the 19th century! I considered extending this statement to all Europeans but there has been quite a bit of change in national borders in Europe over the last 200 years.

The large scale features of the USA arise from a number of sources. The earliest of these originate from the French and Indian War in the mid-18th century which saw the England and the colonists take the territory around the Great Lakes from the French and subsequently take further land from the French in the Louisiana Purchase. Further to the west territory came from the Spanish and then a newly independent Mexico. The border with Canada was agreed largely at the 49th parallel with the British in 1818. Later the Dutch would cede their territory along the Hudson river and the Spanish the last of their territory in what is Florida.

There are some recurring themes determining the shapes of states, one that comes up repeatedly is the desire for Congress to create States of equal size, in the West there are sets of states with the same height (3o) and width (7o). This concept extended to access to resources, so the ports on the Great Lakes are shared amongst the surrounding States. A second big driving force is slavery, the Missouri Compromise placed a boundary at a latitude of 36o 30′ below which slavery was allowed, and above which it was not. This motivated boundaries of states, and led to a battle to create equal numbers of states above and below the line.

There are irregularities. Boston Corner looks like it should belong in  Massachusetts but is actually in New York state, this is because the terrain made access to Boston Corner from the rest of Massachusetts difficult. In the early days this type of inaccessibility led to lawlessness, so states were willing to cede territory to avoid it. Whole states were created to address potential lawlessness, when gold was discovered in what is now Idaho it was felt too distant from Oregon to be ruled from there with the influx of unruly gold miners. There was also a concern that they would displace the coastal Oregonians from government.

Sometimes a river makes a good boundary although when the river has tributaries things get a bit tricky, it is even worse when borders are defined with reference to “head waters” which are notoriously difficult to locate. The other problem with rivers is that they meander – meaning that chunks of a State may find themselves on the “wrong” side of a river when the river moves. In some cases surveying errors and mistakes in negotiations led to oddly formed borders.

The supersize California and Texas states are a result of their own origins in virtual nationhood. Texas was, for a brief period, an independent country which subsequently joined the Union. California formed with the influx of the miners who came for gold, the Union was more concerned that they join than try to enforce borders upon the new State.

The charters of the original US colonies which later evolved into states typically gave them territories that stretched all the way from the Atlantic to the Pacific coast, during the 17th and 18th centuries this was largely moot – colonies scarcely had the wherewithal to maintain small populations on the  Eastern seaboard. The British monarchs granting these charters were not necessarily consistent, or particularly well-advised. So some boundaries are defined by “headwaters” which are notoriously ill-defined.

It is inevitable that the book is a bit repetitive, after all every border has two sides. This is occasionally jarring but usually handled quite well with cross referencing.

Missing from this book is much reference to the Native Americans, they are mentioned as an aside in a few places but little more than that. There is another book in the territories of the Native Americans prior to the European colonisation of the country – I just don’t know where it is! This article on The best books on Native Americans and Colonisers looks like a good place to start.

Overall I quite enjoyed this book, I read most of it on a long train ride. I suspect maps and boundaries are a bit of a niche interest but I feel I also picked up the broad shape of the creation of the USA.

Jan 14 2020

Book review: You look like a thing and I love you by Janelle Shane

You look like a thing and I love you by Janelle Shane is a non-technical overview of machine learning. This isn’t to say it doesn’t goYou look like a thing and I love you book cover into some depth, and that if you are experienced practitioner in machine learning you won’t learn something. The book is subtitled “How Artificial Intelligence Works and Why It’s Making the World a Weirder Place” but Shane makes clear at the outset that it is all about machine learning – Artificial Intelligence is essentially the non-specialist term for the field.

Machine learning is based around training an algorithm with a set of data which represents the task at hand. It might be a list of names (of kittens, for example) where essentially we are telling the algorithm “all these things here are examples of what we want”. Or it might be a set of images where we indicate the presence of dogs, cats or whatever we are interested in. Or, to use one of Shane’s examples, it might be sandwich recipes labelled as “tasty” or “not so tasty”.  After training, the algorithm will be able to generate names consistent with the training set, label images as containing cats or dogs or tell you whether a sandwich is potentially tasty.

The book has grown out of Shane’s blog AI Weirdness where she began posting about her experiences of training recurrent neural networks (a machine learning algorithm) at the beginning of 2016. This started with her attempts to generate recipes. The results are, at times, hysterically funny. Following attempts at recipes she went on to the naming of things, using neural networks to generate the names of kittens, guinea pigs, craft beers, Star Wars planet names and to generate knitting patterns. More recently she has been looking at image labelling using machine learning, and at image generation using generative adversarial networks.

The “happy path” of machine learning is interrupted by a wide range of bumps in the road which Shane identifies, these include:

  • Messy training data – the recipe data, at one point, had ISBN numbers mixed in which led to the neural network erroneously trying to include ISBN-like numbers in recipes;
  • Biased training data – someone tried to analyse the sentiment of restaurant reviews but found that Mexican restaurants were penalised because the Word2vec training set (word2vec is a popular machine learning library which they used in there system) associated Mexican with “illegal”;
  • Not detecting the thing you thought it was detecting – Shane uses giraffes as an example, image labelling systems have a tendency to see giraffes where they don’t exist. This is because if you train a system to recognise animals then in all likelihood you will not include pictures with no animals. Therefore show a neural network an image of some fields and trees with no animals in it will likely “see” an animal because, to its knowledge, animals are always found in such scenes. And neural networks just like giraffes;
  • Inappropriate reward functions – you might think you have given your machine learning system an appropriate “reward function” aka a measure for success but is it really the right one? For example the COMPAS system, which recommends whether prisoners in the US should be recommended for parole, was trained using a reward based on re-arrest, not re-offend. Therefore it tended to recommend against parole for black prisoners because they were more likely to be arrested (not because they were more likely to re-offend);
  • “Hacking the Matrix” – in some instances you might train your system in a simulation of the real world, for example if you want to train a robot to walk then rather than trying to build real robots you would build virtual robots and try them out in a simulated environment. The problem comes when your virtual robot works out how to cheat in the simulated environment, for example by exploiting limitations of collision detection to generate energy;
  • Problems unsuited to machine learning – some tasks are not amenable to machine learning solutions. For example, in the recipe generation problem the “memory” of the neural network limits the recipes generated because by the time a neural network has reached the 10th ingredient in a list it has effectively forgotten the first ingredient. Furthermore, once trained in one task, a neural network will “catastrophically forget” how to do that task if it is subsequently trained to do another task – machine learning systems are not generalists;

My favourite of these is “Hacking the matrix” where algorithms discover flaws in the simulations in which they run, or flaws in their reward system, and exploit them for gain. This blog post on AI Weirdness provides some examples, and links to original research.

Some of this is quite concerning, the examples Shane finds are the obvious ones – the flight simulator which found that it could meet the goal of a “minimum force” landing by making the landing force enormous and overflowing the variable that stored the value, making it zero. This is catastrophic from the pilot’s point of view. This would have been a very obvious problem which could be identified without systematic testing. But what if the problem is not so obvious but equally catastrophic when it occurs?

A comment that struck me towards the end of the book was that humans “fake intelligence” with prejudices and stereotypes, it isn’t just machines that use shortcuts when they can.

The book finishes with how Shane sees the future of artificial intelligence, essentially in a recognition that these systems have strengths and weaknesses and that the way forward is to combine artificial and human intelligence.

Definitely worth a read!