Knowledge – Trance-Scripts

Master Algorithms

Pedro Domingos’s The Master Algorithm has Caius wondering about induction and deduction, a distinction that has long puzzled him.

Domingos distinguishes between five main schools, “the five tribes of machine learning,” as he calls them, each having created its own algorithm for helping machines learn. “The main ones,” he writes, “are the symbolists, connectionists, evolutionaries, Bayesians, and analogizers” (51).

Caius notes down what he can gather of each approach:

Symbolists reduce intelligence to symbol manipulation. “They’ve figured out how to incorporate preexisting knowledge into learning,” explains Domingos, “and how to combine different pieces of knowledge on the fly in order to solve new problems. Their master algorithm is inverse deduction, which figures out what knowledge is missing in order to make a deduction go through, and then makes it as general as possible” (52).

Connectionists model intelligence by “reverse-engineering” the operations of the brain. And the brain, they say, is like a forest. Shifting from a symbolist to a connectionist mindset is like moving from a decision tree to a forest. “Each neuron is like a tiny tree, with a prodigious number of roots — the dendrites — and a slender, sinuous trunk — the axon,” writes Domingos. “The brain is a forest of billions of these trees,” he adds, and “Each tree’s branches make connections — synapses — to the roots of thousands of others” (95).

The brain learns, in their view, “by adjusting the strengths of connections between neurons,” says Domingos, “and the crucial problem is figuring out which connections are to blame for which errors and changing them accordingly” (52).

Always, among all of these tribes, the idea that brains and their worlds contain problems that need solving.

The connectionists’ master algorithm is therefore backpropagation, “which compares a system’s output with the desired one and then successively changes the connections in layer after layer of neurons so as to bring the output closer to what it should be” (52).

“From Wood Wide Web to World Wide Web: the layers operate in parallel,” thinks Caius. “As above, so below.”

Evolutionaries, as their name suggests, draw from biology, modeling intelligence on the process of natural selection. “If it made us, it can make anything,” they argue, “and all we need to do is simulate it on the computer” (52).

This they do by way of their own master algorithm, genetic programming, “which mates and evolves computer programs in the same way that nature mates and evolves organisms” (52).

Bayesians, meanwhile, “are concerned above all with uncertainty. All learned knowledge is uncertain, and learning itself is a form of uncertain inference. The problem then becomes how to deal with noisy, incomplete, and even contradictory information without falling apart. The solution is probabilistic inference, and the master algorithm is Bayes’ theorem and its derivatives. Bayes’ theorem tells us how to incorporate new evidence into our beliefs, and probabilistic inference algorithms do that as efficiently as possible” (52-53).

Analogizers equate intelligence with pattern recognition. For them, “the key to learning is recognizing similarities between situations and thereby inferring other similarities. If two patients have similar symptoms, perhaps they have the same disease. The key problem is judging how similar two things are. The analogizers’ master algorithm is the support vector machine, which figures out which experiences to remember and how to combine them to make new predictions” (53).

Reading Domingos’s recitation of the logic of the analogizers’ “weighted k-nearest-neighbor” algorithm — the algorithm commonly used in “recommender systems” — reminds Caius of the reasoning of Vizzini, the Wallace Shawn character in The Princess Bride.

The first problem with nearest-neighbor, as Domingos notes, “is that most attributes are irrelevant.” “Nearest-neighbor is hopelessly confused by irrelevant attributes,” he explains, “because they all contribute to the similarity between examples. With enough irrelevant attributes, accidental similarity in the irrelevant dimensions swamps out meaningful similarity in the important ones, and nearest-neighbor becomes no better than random guessing” (186).

Reality is hyperspatial, hyperdimensional, numberless in its attributes — “and in high dimension,” notes Domingos, “the notion of similarity itself breaks down. Hyperspace is like the Twilight Zone. […]. When nearest-neighbor walks into this topsy-turvy world, it gets hopelessly confused. All examples look equally alike, and at the same time they’re too far from each other to make useful predictions” (187).

After the mid-1990s, attention in the analogizer community shifts from “nearest-neighbor” to “support vector machines,” an alternate similarity-based algorithm designed by Soviet frequentist Vladimir Vapnik.

“We can view what SVMs do with kernels, support vectors, and weights as mapping the data to a higher-dimensional space and finding a maximum-margin hyperplane in that space,” writes Domingos. “For some kernels, the derived space has infinite dimensions, but SVMs are completely unfazed by that. Hyperspace may be the Twilight Zone, but SVMs have figured out how to navigate it” (196).

Domingos’s book was published in 2015. These were the reigning schools of machine learning at the time. The book argues that these five approaches ought to be synthesized — combined into a single algorithm.

And he knew that reinforcement learning would be part of it.

“The real problem in reinforcement learning,” he writes, inviting the reader to suppose themselves “moving along a tunnel, Indiana Jones-like,” “is when you don’t have a map of the territory. Then your only choice is to explore and discover what rewards are where. Sometimes you’ll discover a treasure, and other times you’ll fall into a snake pit. Every time you take an action, you note the immediate reward and the resulting state. That much could be done by supervised learning. But you also update the value of the state you just came from to bring it into line with the value you just observed, namely the reward you got plus the value of the new state you’re in. Of course, that value may not yet be the correct one, but if you wander around doing this for long enough, you’ll eventually settle on the right values for all the states and the corresponding actions. That’s reinforcement learning in a nutshell” (220-221).

Self-learning and attention-based approaches to machine learning arrive on the scene shortly thereafter. Vaswani et al. publish their paper, “Attention Is All You Need,” in 2017.

“Attention Chaud!” reads the to-go lid atop Caius’s coffee.

Domingos hails him with a question: “Are you a rationalist or an empiricist?” (57).

“Rationalists,” says the computer scientist, “believe that the senses deceive and that logical reasoning is the only sure path to knowledge,” whereas “Empiricists believe that all reasoning is fallible and that knowledge must come from observation and experimentation. […]. In computer science, theorists and knowledge engineers are rationalists; hackers and machine learners are empiricists” (57).

Yet Caius is neither a rationalist nor an empiricist. He readily admits each school’s critique of the other. Senses deceive AND reason is fallible. Reality unfolds not as a truth-finding mission but as a dialogue.

Caius agrees with Scottish Enlightenment philosopher David Hume’s critique of induction. As Hume argues, we can never be certain in our assumption that the future will be like the past. If we seek to induce the Not-Yet from the As-Is, then we do so on faith.

Yet inducing the Not-Yet from the As-Is is the game we play. We learn by observing, inducing, and revising continually, ad infinitum, under conditions of uncertainty. Under such conditions, learning is only ever a gamble, a wager made moment by moment, without guarantees. No matter how large our dataset, we ain’t seen nothing yet.

What matters, then, is the faith we exercise in our interaction with the unknown.

Most of today’s successes in machine learning emerge from the connectionists.

“Neural networks’ first big success was in predicting the stock market,” writes Domingos. “Because they could detect small nonlinearities in very noisy data, they beat the linear models then prevalent in finance and their use spread. A typical investment fund would train a separate network for each of a large number of stocks, let the networks pick the most promising ones, and then have human analysts decide which of those to invest in. A few funds, however, went all the way and let the learners themselves buy and sell. Exactly how all these fared is a closely guarded secret, but it’s probably not an accident that machine learners keep disappearing into hedge funds at an alarming rate” (The Master Algorithm, p. 112).

Nowhere in The Master Algorithm does Domingos interrogate his central metaphor of “mastery” and its relationship to conquest, domination, and control. The enemy is always painted in the book as “cancer.” Yet as any good “analogizer” would know, the Master Algorithm that perfectly targets “cancer” is also the Killer App used by the state against those it encodes as its enemies.

One wouldn’t know this, though, from the future as imagined by Domingos. What he imagines instead is a kind of game: a digital future where each of us is a learning machine. “Life is a game between you and the learners that surround you,” writes Domingos.

“You can refuse to play, but then you’ll have to live a twentieth-century life in the twenty-first. Or you can play to win. What model of you do you want the computer to have? And what data can you give it that will produce that model? Those two questions should always be in the back of your mind whenever you interact with a learning algorithm — as they are when you interact with other people” (264).

The Artist-Activist as Hero

Mashinka Firunts Hakopian imagines artists and artist-activists as heroic alternatives to mad scientists. The ones who teach best what we know about ourselves as learning machines.

“Artists, and artist-activists, have introduced new ways of knowing — ways of apprehending how learning machines learn, and what they do with what they know,” writes Hakopian. “In the process, they’ve…initiated learning machines into new ways of doing. They’ve explored the interiors of erstwhile black boxes and rendered them transparent. They’ve visualized algorithmic operations as glass boxes, exhibited in white cubes and public squares. They’ve engaged algorithms as co-creators, and carved pathways for collective authorship of unanticipated texts. Most saliently, artists have shown how we might visualize what is not yet here” (The Institute for Other Intelligences, p. 90).

This is what blooms here in my library: “blueprints and schemata of a forward-dawning futurity” (90).

The Inner Voice That Loves Me

Stretches, relaxes, massages neck and shoulders, gurgles “Yes!,” gets loose. Reads Armenian artist Mashinka Hakopian’s “Algorithmic Counter-Divination.” Converses with Turing and the General Intellect about O-Machines.

Appearing in an issue of Limn magazine on “Ghostwriters,” Hakopian’s essay explores another kind of O-machine: “other machines,” ones powered by community datasets. Trained by her aunt in tasseography, a matrilineally transmitted mode of divination taught and practiced by femme elders “across Armenia, Palestine, Lebanon, and beyond,” where “visual patterns are identified in coffee grounds left at the bottom of a cup, and…interpreted to glean information about the past, present, and future,” Hakopian takes this practice of her ancestors as her key example, presenting O-machines as technologies of ancestral intelligence that support “knowledge systems that are irreducible to computation.”

With O-machines of this sort, she suggests, what matters is the encounter, not the outcome.

In tasseography, for instance, the cup reader’s identification of symbols amid coffee grounds leads not to a simple “answer” to the querent’s questions, writes Hakopian; rather, it catalyzes conversation. “In those encounters, predictions weren’t instantaneously conjured or fixed in advance,” she writes. “Rather, they were collectively articulated and unbounded, prying open pluriversal outcomes in a process of reciprocal exchange.”

While defenders of western technoscience denounce cup reading for its superstition and its witchcraft, Hakopian recalls its place as a counter-practice among Armenian diasporic communities in the wake of the 1915 Armenian Genocide. For those separated from loved ones by traumas of that scale, tasseography takes on the character of what hauntologists like Derrida would call a “messianic” redemptive practice. “To divine the future in this context is a refusal to relinquish its writing to agents of colonial violence,” writes Hakopian. “Divination comes to operate as a tactic of collective survival, affirming futurity in the face of a catastrophic present.” Consulting with the oracle is a way of communing with the dead.

Hakopian contrasts this with the predictive capacities imputed to today’s AI. “We reside in an algo-occultist moment,” she writes, “in which divinatory functions have been ceded to predictive models trained to retrieve necropolitical outcomes.” Necropolitical, she adds, in the sense that algorithmic models “now determine outcomes in the realm of warfare, policing, housing, judicial risk assessment, and beyond.”

“The role once ascribed to ritual experts who interpreted the pronouncements of oracles is now performed by technocratic actors,” writes Hakopian. “These are not diviners rooted in a community and summoning communiqués toward collective survival, but charlatans reading aloud the results of a Ouija session — one whose statements they author with a magnetically manipulated planchette.”

Hakopian’s critique is in that sense consistent with the “deceitful media” school of thought that informs earlier works of hers like The Institute for Other Intelligences. Rather than abjure algorithmic methods altogether, however, Hakopian’s latest work seeks to “turn the annihilatory logic of algorithmic divination against itself.” Since summer of 2023, she’s been training a “multimodal model” to perform tasseography and to output bilingual predictions in Armenian and English.

Hakopian incorporated this model into “Բաժակ Նայող (One Who Looks at the Cup),” a collaborative art installation mounted at several locations in Los Angeles in 2024. The installation features “a purpose-built Armenian diasporan kitchen located in an indeterminate time-space — a re-rendering of the domestic spaces where tasseography customarily takes place,” notes Hakopian. Those who visit the installation receive a cup reading from the model in the form of a printout.

Yet, rather than offer outputs generated live by AI, Hakopian et al.’s installation operates very much in the style of a Mechanical Turk, outputting interpretations scripted in advance by humans. “The model’s only function is to identify visual patterns in a querent’s cup in order to retrieve corresponding texts,” she explains. “This arrangement,” she adds, “declines to cede authorship to an algo-occultist circle of ‘stochastic parrots’ and the diviners who summon them.”

The ”stochastic parrots” reference is an unfortunate one, as it assumes a stochastic cosmology.

I’m reminded of the first thesis from Walter Benjamin’s “Theses on the Philosophy of History,” the one where Benjamin likens historical materialism to that very same precursor to today’s AI: the famous chess-playing device of the eighteenth century known as the Mechanical Turk.

“The story is told of an automaton constructed in such a way that it could play a winning game of chess, answering each move of an opponent with a countermove,” writes Benjamin. “A puppet in Turkish attire and with a hookah in its mouth sat before a chessboard placed on a large table. A system of mirrors created an illusion that this table was transparent from all sides. Actually, a little hunchback who was an expert chess player sat inside and guided the puppet’s hand by means of strings. One can imagine a philosophical counterpart to this device. The puppet called ‘historical materialism’ is to win all the time. It can easily be a match for anyone if it enlists the services of theology, which today, as we know, is wizened and has to keep out of sight.” (Illuminations, p. 253).

Hakopian sees no magic in today’s AI. Those who hype it are to her no more than deceptive practitioners of a kind of “stage magic.” But magic is afoot throughout the history of computing for those who look for it.

Take Turing, for instance. As George Dyson reports, Turing “was nicknamed ‘the alchemist’ in boarding school” (Turing’s Cathedral, p. 244). His mother had “set him up with crucibles, retorts, chemicals, etc., purchased from a French chemist” as a Christmas present in 1924. “I don’t care to find him boiling heaven knows what witches’ brew by the aid of two guttering candles on a naked windowsill,” muttered his housemaster at Sherborne.

Turing’s O-machines achieve a synthesis. The “machine” part of the O-machine is not the oracle. Nor does it automate or replace the oracle. It chats with it.

Something similar is possible in our interactions with platforms like ChatGPT.

O-Machines

In his dissertation, completed in 1938, Alan Turing sought “ways to escape the limitations of closed formal systems and purely deterministic machines” (Dyson, Turing’s Cathedral, p. 251) like the kind he’d imagined two years earlier in his landmark essay “On Computable Numbers.” As George Dyson notes, Turing “invoked a new class of machines that proceed deterministically, step by step, but once in a while make nondeterministic leaps, by consulting ‘a kind of oracle as it were’” (252).

“We shall not go any further into the nature of this oracle,” wrote Turing, “apart from saying that it cannot be a machine.” But, he adds, “With the help of the oracle we could form a new kind of machine (call them O-machines)” (“Systems of Logic Based on Ordinals,” pp. 172-173).

James Bridle pursues this idea in his book Ways of Being.

“Ever since the development of digital computers,” writes Bridle, “we have shaped the world in their image. In particular, they have shaped our idea of truth and knowledge as being that which is calculable. Only that which is calculable is knowable, and so our ability to think with machines beyond our own experience, to imagine other ways of being with and alongside them, is desperately limited. This fundamentalist faith in computability is both violent and destructive: it bullies into little boxes what it can and erases what it can’t. In economics, it attributes value only to what it can count; in the social sciences it recognizes only what it can map and represent; in psychology it gives meaning only to our own experience and denies that of unknowable, incalculable others. It brutalizes the world, while blinding us to what we don’t even realize we don’t know” (177).

“Yet at the very birth of computation,” he adds, “an entirely different kind of thinking was envisaged, and immediately set aside: one in which an unknowable other is always present, waiting to be consulted, outside the boundaries of the established system. Turing’s o-machine, the oracle, is precisely that which allows us to see what we don’t know, to recognize our own ignorance, as Socrates did at Delphi” (177).

Tuesday October 31, 2017

By traveling through mental circuitry, we open new avenues of action. We come back, and we’re not the same. We possess abstract shape patterns, like afterimages of fireworks. We feel out of it sometimes, asleep to our own reality. I don’t have to think; I just know things now, some of it audio-visual, like materials filed in boxes beneath the floorboards — the Id, the unconscious. But with this new knowledge comes the paranoid sensation that I’m still missing some crucial bit of knowledge possessed by others. Sure, one grants, as Jameson puts it, “the historicity of perception (and of the apparatuses in which it is registered, and registers, all at once)” (Signatures of the Visible, p. 3). But what then becomes of the Imagination? What is imagination’s relation to those historically generated forms known as the sonic, the visual, and the linguistic? Literary scholar R.A. Durr took the terms “psychedelic” and “imaginative” to refer to a “fundamentally identical power of apprehension, or mode of being” (Poetic Vision and the Psychedelic Experience, pp. viii-ix). This imaginative power sends and receives codes. Like in a dream, a long dream. Prove it one can’t. One has to just trust it. The truth is a scenario that becomes by whatever means necessary. Sometimes, however, in order to work I need to perambulate. Get up and move, shake a leg. Jameson thinks absorption in the image results in the negation of thought. Let’s drink to that. Or let’s drink, at least, to the negation of those habitual forms of reason and judgment that dominate life in our time. Let’s drink as well to taking down “The Man.” This was the Yippies’ term for the present system of government. Capitalism has hollowed out reality. It’s been growing and spreading beneath us like a cancer. The characters will make the leap and defeat it, I imagine. But the author? One mustn’t say.