Blaise Aguera y Arcas
Michelangelo "Every block of stone has a statue inside of it, and the job of the sculptor is to discover it." So I think that what Michelangelo was getting at is that we create by perceiving, and that perception itself is an act of imagination and is the stuff of creativity.
The first person who developed some kind of insight into what was going on in the brain was the great Spanish neuroanatomist, Santiago Ramón y Cajal, in the 19th century.
5.29 a bird
6.51 scare quotes
9.47 minimize the error
https://www.ted.com/talks/blaise_aguera_y_arcas_how_computers_are_learning_to_be_creative
Correlation causation
Machine learning does not distinguish between correlations that are causally meaningful and ones that are incidental.
Wu and Zhang are able to use a variety of techniques to explore this in detail. This is especially tractable for the simpler machine learning approaches that involve measuring relationships between standard facial landmarks. They summarize,
“[…] the angle θ from nose tip to two mouth corners is on average 19.6% smaller for criminals than for non-criminals and has a larger variance. Also, the upper lip curvature ρ is on average 23.4% larger for criminals than for noncriminals. On the other hand, the distance d between two eye inner corners for criminals is slightly narrower (5.6%) than for non-criminals.” [7]
We may be able to get an intuitive sense of what this looks like by comparing the top row of “criminal” examples with the bottom row of “non-criminal” examples, shown in the paper’s Figure 1:
Figure 3. Wu and Zhang’s “criminal” images (top) and “non-criminal” images (bottom). In the top images, the people are frowning. In the bottom, they are not. These types of superficial differences can be picked up by a deep learning system.
Darwin wrote in his 1871 book The Descent of Man:
“[…] man bears in his bodily structure clear traces of his descent from some lower form; […] [n]or is the difference slight in moral disposition between a barbarian, such as the man described by the old navigator Byron, who dashed his child on the rocks for dropping a basket of sea-urchins, and a Howard or Clarkson; and in intellect, between a savage who does not use any abstract terms, and a Newton or Shakspeare. Differences of this kind between the highest men of the highest races and the lowest savages, are connected by the finest gradations.”
Unsurprisingly, Darwin’s apex of humanity is peopled by the physicist Isaac Newton, the playwright William Shakespeare, the abolitionist Thomas Clarkson, and the philanthropist John Howard: all were English, Christian, white, male, and from the educated classes — that is, much like Darwin himself. Darwin’s views were in step with (and, in some ways, more progressive than) those of his peers; more generally, they illustrate homophily, the pervasive cognitive bias causing people to identify with and prefer people similar to themselves.An Israeli startup, Faception, has already taken the logical next step, though they have not published any details about their methods, sources of training data, or quantitative results:
“Faception is first-to-technology and first-to-market with proprietary computer vision and machine learning technology for profiling people and revealing their personality based only on their facial image.”
The Faception team are not shy about promoting applications of their technology, offering specialized engines for recognizing “High IQ”, “White-Collar Offender”, “Pedophile”, and “Terrorist” from a face image. [16] Their main clients are in homeland security and public safety.
Dorothea Lange’s famous Depression-era photos, such as her “Migrant Mother” series from 1936, take as their subject the emotional shaping of the human face and body by a difficult environment.
From Dorothea Lange’s “Migrant Mother” series. The original caption reads: “Destitute peapickers in California; a 32 year old mother of seven children. February 1936.”
Perhaps unsurprisingly, the present-day researchers whose work on social perception of faces Wu and Zhang cite as an inspiration tend to take a more nuanced view of the phenomena they are studying. On one hand, this work has shown that people can form character impressions such as trustworthiness from facial appearance after seeing a face for less than one tenth of a second and that these impressions predict important social outcomes, ranging from political elections to economic transactions to legal decisions. On the other hand, while we form impressions almost reflexively from facial appearance, this does not imply that these impressions are accurate. The evidence suggests that they are not.
Fundamentally, the idea that there might be some “criminal type”, and that this is evident on a person’s face, rests on several flawed assumptions:
The appearance of a person’s face is purely a function of innate properties;
“Criminality” is an innate property in a certain group of people;
Criminal judgment by a legal system reliably determines “criminality” in a way that is unaffected by facial appearance.
Most people see the face on the left in the top row as more attractive than the face on the right. Most people also see the face on the left in the bottom row as less attractive than the face on the right. However, the two faces on the left are different images of the same person; so are the two faces on the right.
As Stephen Jay Gould put it in his 1981 book The Mismeasure of Man,
“The spirit of Plato dies hard. We have been unable to escape the philosophical tradition that what we can see and measure in the world is merely the superficial and imperfect representation of an underlying reality. […] The technique of correlation has been particularly subject to such misuse because it seems to provide a path for inferences about causality (and indeed it does, sometimes — but only sometimes).”
While the US comprises about 5% of the world’s population, it contains about 25% of the global prison population — 2.4 million people. Those incarcerated are disproportionately poor and of color; in the US, being a black male makes you nearly seven times likelier to be incarcerated than if you were a white male. [21] This would make a race detector for face images a fairly effective predictor of “criminality” in the US, if by this word we mean — as Wu and Zhang do in China — someone who has been convicted by the legal system.
Are such convictions fair? Due to the long shadow of slavery and systematic discrimination, a disproportionate number of black people in the US live in difficult economic circumstances, and this in itself is associated with increased criminal conviction, as was the case for England’s white economic underclass in the 19th century. However, the incarceration disparity is far greater than one would expect from this effect alone.
“Predictive policing” (listed as one of TIME Magazine’s 50 best inventions of 2011) is an early example of such a feedback loop. The idea is to use machine learning to allocate police resources to likely crime spots. Believing in machine learning’s objectivity, several US states implemented this policing approach. However, many noticed that the system was learning from previous data. If police were patrolling black neighborhoods more than white neighborhoods, this would lead to more arrests of black people; the system then learns that arrests are more likely in black neighborhoods, leading to reinforcement of the original human bias. It does not result in optimal policing with respect to actual incidence of crime.
see also:
rape culture http://styleisviolence.com/rape-culture/