Posts Tagged ‘2001’
“[They] would think that the truth is nothing but the shadows cast by the artifacts.”*…
How do AI models “understand” and represent reality? Is the inside of a vision model at all like a language model? As Ben Brubaker reports, researchers argue that as the models grow more powerful, they may be converging toward a singular “Platonic” way to represent the world…
Read a story about dogs, and you may remember it the next time you see one bounding through a park. That’s only possible because you have a unified concept of “dog” that isn’t tied to words or images alone. Bulldog or border collie, barking or getting its belly rubbed, a dog can be many things while still remaining a dog.
Artificial intelligence systems aren’t always so lucky. These systems learn by ingesting vast troves of data in a process called training. Often, that data is all of the same type — text for language models, images for computer vision systems, and more exotic kinds of data for systems designed to predict the odor of molecules or the structure of proteins. So to what extent do language models and vision models have a shared understanding of dogs?
Researchers investigate such questions by peering inside AI systems and studying how they represent scenes and sentences. A growing body of research has found that different AI models can develop similar representations, even if they’re trained using different datasets or entirely different data types. What’s more, a few studies have suggested that those representations are growing more similar as models grow more capable. In a 2024 paper, four AI researchers at the Massachusetts Institute of Technology argued that these hints of convergence are no fluke. Their idea, dubbed the Platonic representation hypothesis, has inspired a lively debate among researchers and a slew of follow-up work.
The team’s hypothesis gets its name from a 2,400-year-old allegory by the Greek philosopher Plato. In it, prisoners trapped inside a cave perceive the world only through shadows cast by outside objects. Plato maintained that we’re all like those unfortunate prisoners. The objects we encounter in everyday life, in his view, are pale shadows of ideal “forms” that reside in some transcendent realm beyond the reach of the senses.
The Platonic representation hypothesis is less abstract. In this version of the metaphor, what’s outside the cave is the real world, and it casts machine-readable shadows in the form of streams of data. AI models are the prisoners. The MIT team’s claim is that very different models, exposed only to the data streams, are beginning to converge on a shared “Platonic representation” of the world behind the data.
“Why do the language model and the vision model align? Because they’re both shadows of the same world,” said Phillip Isola, the senior author of the paper.
Not everyone is convinced. One of the main points of contention involves which representations to focus on. You can’t inspect a language model’s internal representation of every conceivable sentence, or a vision model’s representation of every image. So how do you decide which ones are, well, representative? Where do you look for the representations, and how do you compare them across very different models? It’s unlikely that researchers will reach a consensus on the Platonic representation hypothesis anytime soon, but that doesn’t bother Isola.
“Half the community says this is obvious, and the other half says this is obviously wrong,” he said. “We were happy with that response.”…
Read on: “Distinct AI Models Seem To Converge On How They Encode Reality,” from @quantamagazine.bsky.social.
Bracket with: “AGI is here (and I feel fine),” from Robin Sloan and “We Need to Talk About How We Talk About ‘AI’,” from Emily Bender and Nanna Inie.
* from Socrates “Allegory of the Cave,” in Plato’s Republic (Book VII)
###
As we interrogate ideas and Ideas, we might recall that it was on this date that the fictional HAL 9000 computer became operational, according to Arthur C. Clarke’s 2001: A Space Odyssey., in which the artificially-intelligent computer states: “I am a HAL 9000 computer, Production Number 3. I became operational at the HAL Plant in Urbana, Illinois, on January 12, 1997.” (Kubrik’s 1968 movie adaptation put his birthdate in 1992.)
Back to the streets…
Following earlier assays of street signage from all over (e.g., here and here), the rubber finally meets the road itself. Readers, the Toynbee Tile…
Franklin Square, Washington, DC (source)
Since the 1980s, several hundred tiles– all roughly the size of an American state license plate, and all bearing roughly the message above– have been found embedded in the pavement of roads in streets in two dozen major U.S. cities and four South American capitals.
There’s no consensus among scholars of the tiles as to their reference or meaning. It’s pretty widely held that the “Toynbee” reference is to historian Arnold Toynbee, perhaps to a passage (in Experiences):
Human nature presents human minds with a puzzle which they have not yet solved and may never succeed in solving, for all that we can tell. The dichotomy of a human being into ‘soul’ and ‘body’ is not a datum of experience. No one has ever been, or ever met, a living human soul without a body… Someone who accepts – as I myself do, taking it on trust – the present-day scientific account of the Universe may find it impossible to believe that a living creature, once dead, can come to life again; but, if he did entertain this belief, he would be thinking more ‘scientifically’ if he thought in the Christian terms of a psychosomatic resurrection than if he thought in the shamanistic terms of a disembodied spirit.
Others suggest that the tiles allude to Ray Bradbury’s story “The Toynbee Convector,” to Arthur C. Clarke’s story “Jupiter V,” or– perhaps, given the direct 2001 reference, most likely– to Stanley Kubrick’s film (in which, readers will recall, hibernating astronauts who had secret training were to be revived upon arrival on Jupiter).
And while there’s no agreement on the identity of the tiler, a majority of enthusiasts believe that “he” is from Philadelphia– both because the City of Brotherly Love hosts the highest concentration of the plaques and because a collection of tiles found there deviate from the norm to ascribe a plot to John S. Knight (of Knight-Ridder, the erst-while newspaper publishers), the Mafia, and others.
See a (nearly) complete list of tiles and their locations here, a set of photos here, and learn how they are implanted here. Visit this site for a peek at a Sundance award-winning documentary on the Tiles.
UPDATE: Further to earlier posts on Lorem Ipsum and it’s bastard children, Bacon Ipsum and Hipster Ipsum, more grievous greeking: Velo Ipsum (for bicycling enthusiasts), and for the reportorially-inclined, Journo Ipsum.
As we watch where we’re walking, we might recall that it was on this date in 1504 that Michelangelo’s 17-foot-tall marble David was unveiled in a public square outside the Palazzo della Signoria, the seat of civic government in Florence.
All together now!…
The song “Daisy Bell” (aka, “Daisy, Daisy”), written by Harry Dacre in 1892, was famously used in 1961 by Max Mathews, John Kelly, and Carol Lockbaumas as the first example of computer-synthesized voice (an accomplishment to which Stanley Kubrick paid homage in 1968 when he had HAL sing “Daisy Bell” at the end of 2001, A Space Odyssey).
Now Aaron Koblin and Daniel Massey have created the Web 2.0 version: Bicycle Built For 2,000… Koblin and Massey used Amazon’s Mechanical Turk web service, prompting workers (folks from 71 different countries, who were paid $ .06 per pop) to listen to a short sound clip, then record themselves imitating what they heard. Koblin and Massey then synthesized the submissions…
… which readers can hear the results here.
As we clear our throats, we might cast a wistful thought back to this date in 1852, when Henry Wells and William G. Fargo joined with several other investors to launch their eponymously-named cross-country freight business. The California gold rush had created an explosive new need, which Wells, Fargo and other “pony express” and stage lines leapt to meet. It was after the Civil War, in 1866, when Wells, Fargo acquired many of their competitors, that it became the dominant supplier. (Ever flexible, they adapted again three years later, when the transcontinental railroad was finished.)
From it’s earliest days, it also functioned as a bank, factoring the shipments of gold that it carried. Indeed, when Wells, Fargo exited the freight business as a result of government nationalization of freight during World War I, the bank (which merged with Nevada National in the first of a series of “transformative transactions”) continued to operate as “Wells, Fargo,” as indeed it does (albeit under unrecognizably evolved ownership) today.
It remains to be seen whether Wells, Fargo will exit the increasingly nationalized banking business.
You must be logged in to post a comment.