From Skynet and the Terminator franchise, through Wargames and Ava in Ex Machina, artificial intelligences pervade our cinematic experiences. But AIs are already in the real world, answering our questions on our phones and making diagnoses about our health. Adam Rutherford asks if we are ready for AI, when fiction becomes reality, and we create thinking machines.
Tagged with “cognition” (16)
YANSS 090 – Questioning the nature of reality with cognitive scientist Donald Hoffman – You Are Not So Smart
Back in the early 1900s, the German biologist Jakob Johann Baron von Uexküll couldn’t shake the implication that the inner lives of animals like jellyfish and sea urchins must be radically different from those of humans.
Uexküll was fascinated by how meaty, squishy nervous systems gave rise to perception. Noting that the sense organs of sea creatures and arachnids could perceive things that ours could not, he realized that giant portions of reality must therefore be missing from their subjective experiences, which suggested that the same was true of us. In other words, most ticks can’t enjoy an Andrew Lloyd Webber musical because, among other reasons, they don’t have eyes. On the other hand, unlike ticks, most humans can’t smell butyric acid wafting on the breeze, and so no matter where you sit in the audience, smell isn’t an essential (or intended) element of a Broadway performance of Cats.
Uexküll imagined that each animal’s subjective experience was confined to a private sensory world he called an umwelt. Each animal’s umwelt was different, he said, distinctive from that of another animal in the same environment, and each therefore was tuned to take in only a small portion of the total picture. Not that any animal would likely know that, which was Uexküll’s other big idea. Because no organism can perceive the totality of objective reality, each animal likely assumes that what it can perceive is all that can be perceived. Each umwelt is a private universe, fitted to its niche, and the subjective experiences of all of Earth’s creatures are like a sea filled with a panoply of bounded virtual realities floating past one another, each unaware that it is unaware.
Like all ideas, Uexküll’s weren’t completely new. Philosophers had wondered about the differences in subjective and objective reality going back to Plato’s cave (and are still wondering). But even though Uexküll’s ideas weren’t strictly original, he brought them into a new academic silo – biology. In doing so, he generated lines of academic research into neuroscience and the nature of consciousness that are still going today.
For instance, when the philosopher Thomas Nagel famously asked, “What is it like to be a bat?” he thought there was no answer to his question because it would be impossible to think in that way. Bat sonar, he said, is nothing like anything we possess, “and there is no reason to suppose that it is subjectively like anything we can experience or imagine.” All one can do, said Nagel, is imagine what it would be like for a person, like yourself, to be a bat. Imagining what it would be like for a bat to be a bat is impossible. This was part of an overall criticism on the limits of reductionist thinking, and is, of course, still the subject of much debate.
The siblings of these notions appear in the writings of everyone from Timothy Leary with his “reality tunnels” to J.J. Gibson’s “ecological optics” to psychologist Charles Tart and his “consensus trances.” From the Wachowski’s Matrix to Kant’s “noumenon” to Daniel Dennett’s “conscious robots,” we’ve been wondering about these questions for a very long time. You too, I suspect, have stumbled on these problems, asking something along the lines of “do we all see the same colors?” at some point. The answer, by the way, is no.
The assumption in most of these musings is that we humans are unique because we can escape our umwelten. We have reason, philosophy, science, and physics which free us from the prison of our limited human perceptions. We can use tools to extend our senses, to see the background radiation left behind by the big bang or hear the ultrasonic laughter of ticklish mice. Sure, the table seems solid enough when we knock on it, and if you were still trapped in your umwelt, you wouldn’t think otherwise, but now you know it is actually mostly empty space thanks to your understanding of protons and electrons. We assume that more layers of truth reveal themselves to us with each successive paradigm shift.
In this episode of the You Are Not So Smart podcast, we sit down with a scientist who is challenging these assumptions.
Donald Hoffman, a cognitive psychologist at the University of California with a background in artificial intelligence, game theory, and evolutionary biology has developed a new theory of consciousness that, should it prove true, would rearrange our understanding of not only the mind and the brain, but physics itself.
“I agree up to a point,” said Hoffman, “that different organisms are in effectively different perceptual worlds, but where I disagree is that these worlds are seeing different parts of the truth. I don’t think they are seeing the truth at all.”
Hoffman wondered if evolution truly favored veridical minds, so he and his graduate students created computer models of natural selection that included accurate perceptions of reality as a variable.
“We simulated hundreds of thousands of random worlds and put organisms in those worlds that could see all of the truth, part of the truth, or none of the truth,” explained Hoffman. “What we found in our simulations was that organisms that saw reality as-it-is could never outcompete organisms that saw none of reality and were just tuned to fitness, as long as they were of equal complexity.”
The implication, Hoffman said, is that an organism that can see the truth will never be favored by natural selection. This suggests that literally nothing we can conceive of can be said to represent objective reality, not even atoms, molecules, or physical laws. Physics and chemistry are still inside the umwelt. There’s no escape.
“If our perceptual systems evolved by natural selection, then the probability that we see reality as it actually is, in any way, is zero. Precisely zero,” said Hoffman.
Well aware that these ideas come across as woo, Hoffman welcomes challenges from his peers and other interested parties, and in the interview you’ll hear what they’ve said so far and how you can investigate these concepts for yourself.
Also in the show, Hoffman explains his ideas in detail in addition to discussing the bicameral mind, artificial intelligence, and the hard problem of consciousness in this mindbending episode about how we make sense of our world, our existence, and ourselves.
Closing keynote at the UIST and SPLASH conferences, October 2014. Preface:
References to baby-steps towards some of the concepts mentioned:
Dynamic reality (physical responsiveness): - The primary work here is Hiroshi Ishii’s "Radical Atoms": http://tangible.media.mit.edu/project/inform/ - but also relevant are the "Soft Robotics" projects at Harvard: http://softroboticstoolkit.com - and at Otherlab: http://youtube.com/watch?v=gyMowPAJwqo - and some of the more avant-garde corners of material science and 3D printing
Dynamic conversations and presentations: - Ken Perlin’s "Chalktalk" changes daily; here’s a recent demo: http://bit.ly/1x5eCOX
Context-sensitive reading material: - http://worrydream.com/MagicInk/ "Explore-the-model" reading material: - http://worrydream.com/ExplorableExplanations/ - http://worrydream.com/LadderOfAbstraction/ - http://ncase.me/polygons/ - http://redblobgames.com/pathfinding/a-star/introduction.html - http://earthprimer.com/
Evidence-backed models: - http://worrydream.com/TenBrighterIdeas/
Direct-manipulation dynamic authoring: - http://worrydream.com/StopDrawingDeadFish/ - http://worrydream.com/DrawingDynamicVisualizationsTalk/ - http://tobyschachman.com/Shadershop/ Modes of understanding: - Jero…
This week we’re joined by Tal Ater to talk about the offline first revolution, the use of service workers, and how UpUp is helping on that front. We also discussed speech recognition and the work he’s doing with annyang. We rounded off the end of the show talking about promoting open source and Tal’s newest project The Open Sourcer.
Are we nearing the singularity, the point where philosophers say the computer programs we create will be smarter than us?
Artificial intelligence is all around us. In phones, in cars, in our homes. Voice recognition systems, predicative algorithms, GPS. Sometimes they may not work very well, but they are improving all the time, you might even say they are learning.
Come on an entertaining journey through the ethics of artificial intelligence or AI, the science behind intelligent computer programs and robotics. Some software engineers think about the philosophy of the artificial intelligence they are creating, others really don’t care.
You’ll also hear two very human AI stories; the strange tale of the robotic resurrection of science fiction author Philip K Dick, and the resulting android that made headlines around the world, an Australian woman and Karen Jacobsen who has a talent for making voice systems warm and human-like, bringing her International fame as the voice of Siri and giving the Aussie accent some much needed street cred.
You know her well - she’s in your phone, in your car and in your home. Is the singularity nigh?
Brain science podcast and interview with Maryanne Wolf - how the brain processes language.
We spy on the new culture of surveillance. Kurt Andersen talks to technologist and philosopher Jaron Lanier about why we have to watch the watchers. An artist meticulously tracks government spy satellites crossing the night sky. A computer scientist explains what goes into building a facial recognition system. And sitting silently in her car, a photographer secretly snaps pictures of strangers in their homes.
In seedy Moscow, ‘coolhunter’ Cayce meets the maker of the footage. Lorelei King concludes the fast-moving thriller.
Searching for footage clues, ‘coolhunter’ Cayce locates the menacing Baranov. Lorelei King reads the fast-moving thriller.
‘Coolhunter’ Cayce Pollard is mugged in Tokyo - but who’s behind her assailants? Fast-moving thriller read by Lorelei King.
Page 1 of 2Older