I spent near on ten years thinking about automated skin cancer detection. There are various approaches you might use — cyborg human/machine hybrids were my personal favourite — but we settled on more standard machine learning approaches. Conceptually what you need is straightforward: data to learn from, and ways to lever the historical data to the future examples. The following quote is apposite.

One is that, for all the advances in machine learning, machines are still not very good at learning. Most humans need a few dozen hours to master driving. Waymo’s cars have had over 10m miles of practice, and still fall short. And once humans have learned to drive, even on the easy streets of Phoenix, they can, with a little effort, apply that knowledge anywhere, rapidly learning to adapt their skills to rush-hour Bangkok or a gravel-track in rural Greece.

Driverless cars are stuck in a jam – Autonomous vehicles

You see exactly the same thing with skin cancer. With a relatively small number of examples, you can train (human) novices to be much better than most doctors. By contrast, with the machines you need literally hundreds and thousands of examples. Even when you start with large databases, as you parse the diagnostic groups, you quickly find out that for many ‘types’ you have only a few examples to learn from. The rate limiting factor becomes acquiring mega-databases cheaply. The best way to do this is to change data acquisition from a ‘research task’ to a matter of grabbing data that was collected routinely for other purposes (there is a lot of money in digital waste — ask Google).

Noam Chomsky had a few statements germane to this and much else that gets in the way of such goals (1).

Plato’s problem: How can we know so much when the evidence is do slight.

Orwell’s problem: How do we remain so ignorant when the evidence is so overwhelming.

(1): Noam Chomsky: Ideas and Ideals, Cambridge University Press, (1999). Neil Smith.

Books

We commonly think of books as containers of ideas or wrapping for literature, but they can be understood in other ways—as if they were blood cells carrying oxygen through a body politic or data points as infinite as stars in the sky. Books lead lives of their own, and they intersect with our lives in ways we have only begun to understand.

Best Sellers by the Bargeload | by Robert Darnton | The New York Review of Books

A job for life

In the essay “Telling,” he describes the upsetting case of the director of a hospital who, struck down by Alzheimer’s, is admitted to his own hospital. He behaves as if he were still running it, until one day by chance he picks up his own chart. “That’s me,” he says, recognizing his name on the cover. Inside, he reads “Alzheimer’s disease” and weeps. In the same hospital a former janitor is admitted; he too is convinced that he is still working there. He is given harmless tasks to perform; one day he dies of a sudden heart attack “without perhaps ever realising that he had been anything but a janitor with a lifetime of loyal work behind him.”

Truth, Beauty, and Oliver Sacks | by Simon Callow | The New York Review of Books

My mother, a nurse, took on such imagined roles when she too was demented and in a care home.

Obituaries are a source of much joy and enlightenment. None more so than those in the Economist. Last week’s was devoted to the ’60’s photographer Terry O’Neill (you can see some of his iconic images here.

Stars had been his subject since 1962, when he was sent to photograph a new band at the Abbey Road Studios. The older blokes at the Sketch scorned that kind of work, but the young were clearly on the rise, and he was by far the youngest photographer in Fleet Street at the time. At the studios, to get a better light, he took the group outside to snap them holding their guitars a bit defensively: John, Paul, George and Ringo. Next day’s Sketch was sold out, and he suddenly found himself with the run of London and all the coming bands, free to be as creative as he liked. A working-class kid from Romford whose prospects had been either the priesthood or a job in the Dagenham car plant, like his dad, had the world at his feet. He wouldn’t have had a prayer, he thought, in any other era.

And obviously it couldn’t last. In a couple of years he would find a proper job, as both the Beatles and the Stones told him they were going to. For it was hardly serious work to point your Leica at someone and go snap, snap.

Obituary: Terry O’Neill died on November 16th – Catching the moment

The reason I found this particularly interesting is the way social mobility appeared to work and the way it was tied to genuine innovation and social change. I have always loved the trope that when jobs are plentiful, and your committments minimal, you can literally tell the boss to FO on a Friday and start another job on the Monday. Best of all you can experiment and experiment lifts all. This to me is one of the best 1960’s rock n’ roll stories.

If you lift your head above the parapet in universities you come across various conventional wisdoms. One relates to ‘mental wellbeing’ or ‘mental issues’, and another is the value of education in increasing social mobility. My problem is that in both cases there seem (to me at least) many important questions that remain unanswered. For the former, are we talking about mental illness (as in disease) or something else? How robust is the data — aside from self-reporting? The widely reported comments from the former President of the Royal College of Pyschiatrists receive no answer (at last not in my institution). An example: I have sat in a meeting in which one justification for ‘lecture capture’ (recording of live lectures) was to assist students with ‘mental health issues’. But do they help in this context? Do we trust self-reflection in this area? Under what conditions do we think they help or harm?

Enhancing life chances and social mobility is yet another area that I find difficult. I picked up on a comment from Martin Wolf in the FT

We also believe that changing individual characteristics, principally via education, will increase social mobility. But this is largely untrue. We need to be far more honest.

He was referring to the work of John Goldthorpe in Oxford. Digging just a little beneath the surface made me realise that much of what I had believed may not true. Goldthorpe writes:

However, a significant change has occurred in that while earlier, in what has become known as the golden age of mobility, social ascent predominated over social descent, the experience of upward mobility is now becoming less common and that of downward mobility more common. In this sense, young people today face less favourable mobility prospects than did their parents or their grandparents.

This research indicates that the only recent change of note is that the rising rates of upward, absolute mobility of the middle decades of the last century have levelled out. Relative rates have remained more or less constant back to the interwar years. According to this alternative view, what can be achieved through education, whether in regard to absolute or relative mobility, appears limited.

[Jnl Soc. Pol. (2013), 42, 3, 431–450 Cambridge University Press 2013 doi:10.1017/S004727941300024X]

There is a witty exchange in Propect between the journalist (JD) and Goldthorpe (JG).

JD: Would you say that this is something that politicians, in particular, tend not to grasp?

JG: Yes. Tony Blair, for instance, was totally confused about this distinction [between absolute and relative rates of mobility]. He couldnʼt see that the only way you can have more upward mobility in a relative perspective is if you have more downward mobility at the same time. I remember being in a discussion in the Cabinet Office when Geoff Mulgan was one of Blairʼs leading advisors. It took a long time to get across to Mulgan the distinction between absolute and relative rates, but in the end he got it. His response was: “The Prime Minister canʼt go to the country on the promise of downward mobility!”

On both these topics I am conflicted. And on both these topics there are the tools that characterize scholarly inquiry to help guide action: this is what universities should be about. I am however left with a strong suspicion that few are interested in digging deep, rather we choose sound bites over understanding. Working in a university often feels like the university must be somewhere else. That is the optimistic version.

Not only God plays dice

There is an article this week in Nature about how some funders are explicitly funding grant proposals randomly (lotteries). The cynic might say they have been doing this for a longtime.

The article is about pharma and the way its interests flit because of perceived commercial rather than clinical value. There are two phrases that should make you sit up.

1. ‘Scientists were given incentives to meet milestones in the clinical trials,’
2. ‘One precondition for success is a large salesforce’.

The first phrase, is scary. We already know how dishonest much of pharma is. We can manage well without more perverse incentives. Short term shareholder value wins over morality every time.

The second begs the question: if the evidence is good, why do you need to flog your medicine with advertising? A collection of data sheets — with citations — is all you need. And since most pharma spends more on advertising than research, here is a simple way to reduce drug costs. (The answer is of course, that advertising sells more than research — shame on us all).

Novartis cholesterol deal highlights mass-market opportunity | Financial Times

I was back in Dublin a few weeks ago for a family celebration. Then last night — on C4 I think— I was listening to an interview with Fintan O’Toole. Something stirred and below are two quotes from Brian Friel: the context may be Brexit, the reality is something much more.

As a character in Translations says, describing his own fading Gaelic world, “a civilisation can be imprisoned in a linguistic contour which no longer matches the landscape of fact”.

 

To remember everything is a form of madness,” warns one of Friel’s characters.

(1) Brexit, the UK and Ireland: a dialogue of the deaf | Financial Times

This is from the Guardian. The background is serious allergic reactions to food components, and allowing accessible information about what purchased food contains. In her phrase, ‘high-profile casualties on the high street’ she is referring to businesses; I am sure others may have read it differently.

But Kate Nicholls, the chief executive of UKHospitality, said a law change could have a serious impact on the viability of some of the 100,000 restaurants her organisation represents. “Hospitality and particularly high street restaurants are under intense cost pressures and are struggling,” she said. We’ve had a number of high-profile casualties on the high street. Those businesses operate on tight net profit margins. And there’s no doubt some would not be able to cope with any significant change in their cost structure.”

(BTW: she thinks ‘training’ is the solution. Training and education are offered as the answer to everything…”education, education, education”. If only.)

Grieving family’s call for allergy law gets cool response | Society | The Guardian

Carbon offsetting is shaping up to be the greatest mis-selling scandal since the Dominican friar Johann Tetzel sold pardons to redeem the dead. Martin Luther attacked this practice in 1517, in his 95 theses.

Five hundred years later, those of us who seek planetary redemption should reduce our carbon footprint in ways that we control — rather than relying on middlemen who may or may not plant trees. The road to hell, I seem to remember, was paved with good intentions.

Well, the Catholic church usually got there first.

Carbon offset gold rush is distracting us from climate change | Financial Times

Frank Davidoff had a telling phrase about clinical expertise. He likened it to “Dark Matter”. Dark Matter makes up most of the universe, but we know very little about it. In the clinical arena I have spent a lot of time reading and thinking about ‘expertise’,  without developing any grand unifying themes of my own worth sharing. But we live in a world where ‘expertise’ in many domains is under assault, and I have no wise thoughts to pull together what is happening. I do however like (as ever) some nice phrases from Paul Graham. I can’t see any roadmap here just perspectives and shadows.

When experts are wrong, it’s often because they’re experts on an earlier version of the world.

any

Instead of trying to point yourself in the right direction, admit you have no idea what the right direction is, and try instead to be super sensitive to the winds of change.

How to Be an Expert in a Changing World