Wednesday, 18 October 2017

Mastering the game of Go without human knowledge (2017)

An incredible paper in Nature on a DeepMind AI breakthrough.


A long-standing goal of artificial intelligence is an algorithm that learns, tabula rasa, superhuman proficiency in challenging domains. Recently, AlphaGo became the first program to defeat a world champion in the game of Go. The tree search in AlphaGo evaluated positions and selected moves using deep neural networks. These neural networks were trained by supervised learning from human expert moves, and by reinforcement learning from self-play. Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules. AlphaGo becomes its own teacher: a neural network is trained to predict AlphaGo’s own move selections and also the winner of AlphaGo’s games. This neural network improves the strength of the tree search, resulting in higher quality move selection and stronger self-play in the next iteration. Starting tabula rasa, our new program AlphaGo Zero achieved superhuman performance, winning 100–0 against the previously published, champion-defeating AlphaGo.

Original paper HERE.
Great write up in The Guardian HERE

The Deeper Roles of Mathematics in Physical Laws (2015)

HERE is a really thought provoking essay by the physicist Kevin H. Knuth on the mystery of why mathematics is seemingly universal in the physical sciences. This has some technical content, but not so much, given how profound an argument Knuth is making. 


Many have wondered how mathematics, which appears to be the result of both human creativity and human discovery, can possibly exhibit the degree of success and seemingly-universal applicability to quantifying the physical world as exemplified by the laws of physics. In this essay, I claim that much of the utility of mathematics arises from our choice of description of the physical world coupled with our desire to quantify it. This will be demonstrated in a practical sense by considering one of the most fundamental concepts of mathematics: additivity. This example will be used to show how many physical laws can be derived as constraint equations enforcing relevant symmetries in a sense that is far more fundamental than commonly appreciated.

Below, from his conclusion:

The results here shed light on the long-standing questions surrounding the unreasonable effectiveness of mathematics. We have seen interplay between two different aspects of mathematics. The first aspect is related to ordering and symmetries, and the second aspect is related to quantification and the equations that enable one to quantify things. Our choices in the phenomena that we focus on,the descriptions we adopt and the comparisons that we find important often amount to selecting a particular concept of ordering, which can possesses symmetries. The ordering relation and its symmetries in turn constrain consistent attempts at quantification resulting in constraint equations, which in many cases are related to what are considered to be physical laws. Much of the wonderment surrounding the unreasonable effectiveness of mathematics is not associated with the first aspect of ordering and symmetries since these more clearly depend on a choice of description and comparison, which in turn results in symmetries that can be easily observed and verified. Instead such wonderment is associated with the fact that we have equations that consistently allow us to quantify the physical world, and that these equations not only work very well, but in many cases exhibit some degree of universality. If we consider the equations themselves to be fundamental then the success of mathematics is somewhat of a mystery. But if we step back and release ourselves from familiarity and consider order and symmetry to be fundamental, then we see these equations as rules to constrain our artificial quantifications in accordance with the underlying order and symmetries of our chosen descriptions. 

Tuesday, 17 October 2017

Quantifying is a Committing Task

A slide from a talk I gave in Manchester 2014: `Quantifying is a committing task.' (HERE).

The phrase `Quantifying is a committing task' is taken from Cruz-Orive, L.M. (1994). `Toward a more objective biology'. Neurobiology of Aging, Vol. 15, Iss. 3, pp 377-378

Image Copyright M.Reed 2014.

The Best Way to Make Sense of the World

Here is an entertaining profile of the American mathematician Rebecca Goldin - who has "... made it her life’s work to improve quantitative literacy". 

Image from Components of adaptive variation in pinus contorta from the inland northwest by G.E. Rehfeldt (1987) Intermountain Research Station, Ogden, Utah. HERE.

Sunday, 15 October 2017

Where the Wild Winds Are: Walking Europe's Winds from the Pennines to Provence (2017)

And then one day I saw a map with paths I hadn’t seen before. It was a map of Europe transfigured by coloured lines, marauding arrows like troop advances that ploughed across borders, over land and sea, connecting regions and cultures that seemed quite separate in my mind: Latin with Slavic, continental with coastal, North African with southern European. These mysterious corridors had names every bit as tantalising as the Silk Road or the Camino de Santiago: the Mistral, the Tramontana, the Foehn, the Sirocco, the Bora. There was even one in the north of England, more brusquely named the Helm. The map showed the routes of local winds, which blow with tremendous force at specific times of year – normally at the transitions between seasons, such as when winter turns to spring – and, I was intrigued to discover, they were said to influence everything from architecture to psychology. The fact that these invisible powers had names, rather than simply compass directions that described where they were from, gave them a sense of majesty, even of personality. They sounded like characters I could meet. Those swooping, plunging arrows suggested routes I might follow, trails that had not been walked before. As soon as I saw that map I knew: I would follow the winds.

Map of Europe’s named winds by Rodney Paull, 2017


The Psychology of Good Judgement (1996)

Another of Gerd Gigerenzer's papers on both frequency formats and fast and frugal decision making (HERE). 

The Psychology of Good Judgement
Frequency Formats and Simple Algorithms

Gerd Gigerenzer, PhD


Mind and environment evolve in tandem—almost a platitude. Much of judgement and decision making research, however, has compared cognition to standard statistical models, rather than to how well it is adapted to its environment. The author argues two points. First, cognitive algorithms are tuned to certain information formats, most likely to those that humans have encountered during their evolutionary history. In particular, Bayesian computations are simpler when the information is in a frequency format than when it is in a probability format. The author investigates whether frequency formats can make physicians reason more often the Bayesian way. Second, cognitive algorithms need to operate under constraints of limited time, knowledge, and computational power, and they need to exploit the structures of their environments. The author describes a fast and frugal algorithm, Take The Best, that violates standard principles of rational inference but can be as accurate as sophisticated "optimal" models for diagnostic inference. 

Key words: Bayes' theorem; bounded rationality; information format; probabilistic reasoning; satisficing; training; medical education. 

How to Improve Bayesian Reasoning Without Instruction: Frequency Formats (1995)

A paper HERE by Gerd Gigerenzer and Ulrich Hoffrage that argues that natural frequencies are a much better way of presenting, examining and discussing arguments based on Bayes Theorem than other approaches. In the paper, Gigerenzer cites Laplace's assertion that “the theory of probability is at bottom nothing more than good sense reduced to a calculus which evaluates that which good minds know by a sort of instinct, without being able to explain how with precision.” 

The Prosecutor's fallacy (1987)

Interpretation of statistical evidence in criminal trials: The prosecutor's fallacy and the defense attorney's fallacy.

Thompson, William C.,Schumann, Edward L.
Law and Human Behavior, Vol 11(3), Sep 1987, pp 167-187

Two experiments tested 217 undergraduates' ability to use statistical evidence on the incidence rate of a "matching" characteristic when judging the probable guilt of a criminal suspect, based on written descriptions of evidence. Exp I varied whether incidence rate statistics were presented as conditional probabilities or as percentages and found the former promoted inferential errors favoring the prosecution, while the latter produced more errors favoring the defense. Exp II exposed Ss to 2 fallacious arguments on how to interpret the statistical evidence. The majority of Ss failed to detect the error in one or both of the arguments. In both experiments, a comparison of Ss' judgments to Bayesian norms revealed a general tendency to underutilize statistical evidence. 

Image from HERE.

Doing Cartography as Professional Amateurs (2017)

An example map from an excellent talk, Doing Cartography as Professional Amateurs, by Alan McConchie of Stamen Design, San Francisco (HERE).

Saturday, 14 October 2017

The Spirit of Voltaire and Rousseau (1794)

The “Spirit” of Voltaire and Rousseau leading these two celebrated writers to the temple of glory and immortality Anonymous etching, ca. 1794

Image from HERE.

Thursday, 12 October 2017

A Few Remarks on the Helm Wind (1836)

A few remarks on the Helm Wind. By the Rev. William Walton, of Allenheads, near Hexham. Communicated by P. M. Roget, M.D., Sec. U.S. 

On the western declivity of a range of mountains, extending from Brampton, in Cumberland, to Brough, in Westmoreland, a distance of 40 miles, a remarkably violent wind occasionally prevails, blowing with tremendous violence down the western slope of the mountain, extending two or three miles over the plain at the base, often overturning horses with carriages, and producing much damage, especially during the period when ripe corn is standing. It is accompanied by a loud noise, like the roaring of distant thunder: and is carefully avoided by travellers in that district, as being fraught with considerable danger. It is termed the helm wind; and its presence is indicated by a belt of clouds, denominated the helm bar, which rests in front of the mountain, three or four miles west of its summit, and apparently at an equal elevation, remaining immoveable during twenty-four or even thirty-six hours, and collecting or attracting to itself all the light clouds which approach it. As long as this bar continues unbroken, the wind blows with unceasing fury, not in gusts, like other storms, but with continued pressure. This wind extends only as far as the spot where the bar is vertical, or immediately over head; while at the distance of a mile farther west, as well as to the east of the summit of the mountain, it is not unfrequently almost a perfect calm. The author details the particulars of an expedition which lie made with a view to investigate the circumstances of this remarkable meteorological phenomenon, and proposes a theory for its explanation. 

A Meteorological Journal kept at Allenheads, 1400 feet above the level of the Sea, from the 1st of May to the 1st of November, 1836. By the Rev. William Walton. Communicated by P. M. Roget, M.D., Sec. R.S. 

Monday, 9 October 2017

The Mind of John McPhee (2017)

Recently, the New York Times magazine published this superb profile of John McPhee (HERE). 

Sunday, 1 October 2017


The Super-Kamiokande apparatus!

Grand Unification Dream Kept at Bay (2017)

Great piece HERE on how careful high quality observation (not experiment) by Japanese physicists at the Super Kamiokande over a 20 year period have not seen proton decay - and not supported some of the most cherished Grand Unified Theories of theoretical physicists. 

Oh dear.

Q: How could it be that nature isn't as mathematically beautiful as thousands of mathematical physicists willed that it must be so? 
A: The Mind Projection Fallacy strikes again.  


Another story to file under the heading of Thomas Henry Huxley's assertion: 

The great tragedy of Science - the slaying of a beautiful hypothesis by an ugly fact.  

Thursday, 28 September 2017

The American Annual of Photography (1911)

From HERE.

Inside the Painters Studio (2009)

Inspiration is for amateurs — the rest of us just show up and get to work. And the belief that things will grow out of the activity itself and that you will — through work — bump into other possibilities and kick open other doors that you would never have dreamt of if you were just sitting around looking for a great ‘art idea.’ And the belief that process, in a sense, is liberating and that you don’t have to reinvent the wheel every day. Today, you know what you’ll do, you could be doing what you were doing yesterday, and tomorrow you are gonna do what you did today, and at least for a certain period of time you can just work. If you hang in there, you will get somewhere.

From an interview with Chuck Close by Joe Fig in his book  - Inside the Painters Studio

How to fit an elephant (1953)

A superb story HERE by Freeman Dyson about how a meeting with Enrico Fermi changed his career. At the heart of the story, Fermi quotes John von Neumann, who famously said:

With four parameters I can fit an elephant, and with five I can make him wiggle his trunk.

An example of how to actually do this, with Python code is HERE. 

On the Disposition of Iron in Variegated Strata (1868)

More of these images HERE.

Wednesday, 27 September 2017

On Exactitude in Science (1946)

 By Jorge Luis Borges. More HERE.

Compassion, Empathy, Flapdoodle (2017)

An excoriating, but highly entertaining, review in the Dublin Review of Books by Seamus O’Mahony of three books on empathy and neuro science. HERE.  

O'Mahony is perhaps not the most typical book reviewer - he is a  Consultant Gastroenterologist at Cork University Hospital who also writes on medical humanities.

Monday, 18 September 2017

Peer review: not as old as you might think (2015)


Peer review - is not as old as you might think (HERE). Above the Google ngram for `peer review'.

Best Practices for Scientific Computing (2014)

 A really good, readable paper. HERE.

Saturday, 16 September 2017

The Book on the Floor (2016)

This book begins with an in depth, but entertaining, analysis of the famous picture of Malraux and one of his books spread out in front of him in his salon taken in 1954. HERE.

Wednesday, 13 September 2017

Unendurable Line (2017)

Unendurable Line: A Fun Short Film that tracks the Movement of Everyday Objects as a Real-Time Graph. HERE.

Saturday, 9 September 2017

Draft No. 4 (2017)

Here is Michael Durda's review of John McPhee's new book: Draft No. 4. He ends the review by quoting McPhee:

“Creativity lies in what you choose to write about, how you go about doing it, the arrangement through which you present things, the skill and the touch with which you describe people and succeed in developing them as characters, the rhythms of your prose, the integrity of the composition, the anatomy of the piece (does it get up and walk around on its own?), the extent to which you see and tell the story that exists in your material, and so forth. Creative nonfiction is not making something up but making the most of what you have.”

John McPhee - Structure (2013)

A great piece in the New Yorker by John McPhee on how he structures his long non-fiction pieces. The image is a series of the illustrations he has of his structural motifs. From HERE

Thursday, 7 September 2017

Sunday, 3 September 2017

Harmonics, Patterns, and Dynamics in Roman and Italic type (2016)

The website for Frank Bloklund's PhD Thesis at Leiden University - On the Origin of Patterning in Movable Latin Type. (HERE).

Saturday, 2 September 2017

The Shape of Design - Frank Chimero

Frank Chimero's "odd little design book",  The Shape of Design, is available to read online HERE.

On Margins Podcast (2017)

On Margins is a podcast by Craig Mod on Book making (HERE).

The Data Visualisation Catalogue

The Data Visualisation Catalogue HERE.

Friday, 1 September 2017

To Make a Book, Walk on a Book (2017)

 A great essay on the practicalities of designing a book, by Craig Mod (HERE).

Monday, 28 August 2017

Ichi-F: A Worker’s Graphic Memoir of the Fukushima Nuclear Power Plant (2017)

A piece HERE in the New York Review of Books on Ichi-F: A Worker’s Graphic Memoir of the Fukushima Nuclear Power Plant by Kazuto Tatsuta

Aesops Fables (1912)

From HERE.

Problems, Myths and Stories (1999)

We take stories and story-telling for granted. The great reservoir of myths, legends, parables, tales, that we dip into for entertainment, use for films and plays, refer to so as to elucidate a point or draw a parallel – it is always there and we hardly think about it. Tales are as old as humanity, like a long shadow thrown by our history. How old? We don’t know. Whenever we reach a point where it seems impossible to go back further, then we can be sure that soon the dark of our ignorance will yield to research and, behold, it is evident that the long shadow showed itself much earlier than we had thought.

Problems, Myths and Stories by Doris Lessing, HERE.

Image from HERE.


Sunday, 27 August 2017

The earliest English version of the fables of Bidpai; The morall philosophie of Doni, by Sir Thomas North. (1888)

The first English translation of the ancient sanskrit collection of fables, The Panchatantra, was published in 1570 by Sir Thomas North.  A reprint was made in 1888, which is available HERE.

Friday, 25 August 2017

Ancient Babylonian Math Tables deciphered (2017)

Plimpton 322 is a 3,700-year-old Babylonian clay tablet, that has a sophisticated set of trigonometric tables inscribed in the base 60 mathematical notation that the Babylonians used. It includes extensive use of Pythagoras Theorem and Pythagorean triples (3,4,5 is the simplest but the table includes less obvious triples such as 119, 120 and 169). It has just been fully deciphered. Story in the Guardian HERE

Friday, 18 August 2017

Daedalus RIP (2017)

The obituary HERE of David Jones, the person who wrote countless Daedalus colums for New Scientist, then later Nature and the Guardian.

Thursday, 17 August 2017

Despite limited statistical power (2017)

A great short post by Andrew Gelman HERE.

It’s not always clear what people mean by this expression, but sometimes it seems that they’re making the “What does not kill my statistical significance makes it stronger” fallacy, thinking that the attainment of statistical significance is a particular feat in the context of a noisy study, so that they’re (mistakenly) thinking of the “limited statistical power” of that study as a further point in favor of their argument.

Wednesday, 16 August 2017

Moses and the Brazen Serpent and the Transfiguration of Jesus (1683)

HERE is a fantastic time lapse video of the unpacking and hanging of Cristóbal de Villalpando's Moses and the Brazen Serpent and the Transfiguration of Jesus at the Met. The installation of the 28-foot-tall painting took five days.

Genome re-sequencing reveals the history of apple... (2017)

A paper HERE on the complex history of how the domestic apple developed along the `silk road' trading routes. 

The abstract reads:

Human selection has reshaped crop genomes. Here we report an apple genome variation map generated through genome sequencing of 117 diverse accessions. A comprehensive model of apple speciation and domestication along the Silk Road is proposed based on evidence from diverse genomic analyses. Cultivated apples likely originate from Malus sieversii in Kazakhstan, followed by intensive introgressions from M. sylvestris. M. sieversii in Xinjiang of China turns out to be an “ancient” isolated ecotype not directly contributing to apple domestication. We have identified selective sweeps underlying quantitative trait loci/genes of important fruit quality traits including fruit texture and flavor, and provide evidences supporting a model of apple fruit size evolution comprising two major events with one occurring prior to domestication and the other during domestication. This study outlines the genetic basis of apple domestication and evolution, and provides valuable information for facilitating marker-assisted breeding and apple improvement.

On the Earthshine depicted in Galileo's watercolors of the Moon (1609 / 2017)

A great paper HERE by Paolo Molaro, carefully re-analysing the wash studies that Galileo made of the moon in 1609. The abstract reads:

With the manuscript of the Sidereus Nuncius preserved at the Biblioteca Nazionale of Florence are included 7 watercolors of the Moon painted by Galileo himself. We suggest that some of them, and in particular the drawing of the 30 Nov 1609 of the very first Moon's observations, illustrate the phenomenon of the Earthshine of the Moon, which was discussed in some detail in the Sidereus Nuncius to provide evidence of the similarity of Earth to other celestial bodies. The watercolors were used as models for the engraving of the Moon in the Sidereus but, surprisingly, the secondary light had not been reproduced. Galileo may have decided for the inclusion of the passage on the Earthshine only at a very late stage of the editorial process. Galileo's hesitation shows how contentious was this issue already recognized as a possible discriminant between the different systems of the world.  

Saturday, 5 August 2017

Book of Miracles (1550)

The Book of Miracles is a recently discovered book of apocalyptic Renaissance art. The illustrated manuscript was created in 1550's Augsburg. Taschen have just published an annotated and fully illustrated edition (HERE).

Friday, 4 August 2017

Sir, I exist! (1899)

from War Is Kind (1899) by Stephen Crane (1871-1900)

Friday, 28 July 2017

Example Computer Plots (1980)

From Byte Magazine 1980 (HERE).

Thursday, 27 July 2017

Spontaneous fine-tuning to environment in many-species chemical reaction networks (2017)

HERE is a computer simulation study of how a dilute solution of different `chemical' species can spontaneously begin to capture energy and create non-random structures (cf life).


A chemical mixture that continually absorbs work from its environment may exhibit steady-state chemical concentrations that deviate from their equilibrium values. Such behavior is particularly interesting in a scenario where the environmental work sources are relatively difficult to access, so that only the proper orchestration of many distinct catalytic actors can power the dissipative flux required to maintain a stable, far-from-equilibrium steady state. In this article, we study the dynamics of an in silico chemical network with random connectivity in an environment that makes strong thermodynamic forcing available only to rare combinations of chemical concentrations. We find that the long-time dynamics of such systems are biased toward states that exhibit a fine-tuned extremization of environmental forcing.

Tuesday, 25 July 2017

Voyage d'Urien (1928)

Two images from Voyage d'Urien by André Gide, with engravings by Alfred Latour, set in Lutetia by Jan van Krimpen. Published in 1928 by the Halcyon Press, Maastricht, by A.A.M. Stols. Images from HERE.

Monday, 24 July 2017

Open Letter to an Editor on Wood Engraving (1940)

'Each time one looks at the beautiful highly polished surface of a new piece of boxwood one should set out with the thought that one has the whole of one's life in which to make one perfect engraving, so there is no need to hurry. And in actual fact it is not easy to do so. What with the natural limitations of the wood, the fact that one is working in reverse and with white on black, instead of the normal black on white, there are sufficient difficulties to overcome without imperilling any chance of success by working to time.' 

John Buckland Wright,  Open Letter to an Editor on Wood Engraving (1940). 

Image above, the ex libris that John Buckland Wright created in 1932 for the Dutch collector M.B.B. Nijkerk. Image from HERE.

Sunday, 23 July 2017

Joy and Pain (1980)

One of my all time favourite songs: Joy and Pain by Maze Featuring Frankie Beverly. A great piece on their 40 year career HERE. A live version of the track HERE.

Pi is Wrong (2001)

HERE is the seminal paper from 2001 by Bob Palais on why the mathematical constant pi is wrong. Recent adherents of this paper have created a manifesto dedicated to building support for the alternative to pi - tau.


Related Posts Plugin for WordPress, Blogger...