Is cosmology in crisis?

In past posts, when I’ve written about the expansion of the universe, I’ve generally referred to the rate of that expansion, the Hubble constant, as being around 70 km/s/megaparsec, that is, for every megaparsec a galaxy is distant from us, it’s moving away at 70 kilometers per second faster.  So a galaxy 100 megapasecs away is moving away at 7000 km/s, and one 200 megaparsecs away at 14000 km/s.

But I have to admit this was an oversimplification.  70 km/s was actually a rough and rounded averaging of two measurements for the expansion, one taken using the cosmic distance ladder, and the other using observations of the cosmic background radiation.  The former currently yields about 74 km/s, and the latter about 67 km/s.

Everyone for a long time thought this difference was just a measuring artefact that would eventually be smoothed out.  Everyone is turning out to be wrong.  As this news story discusses, the two measurements have been refined extensively.  Confidence in these individual measurements are pretty high, and the margins of error don’t overlap.

In other words, either one or both of these methods has assumptions in it that are wrong, or there is something completely unexpected going on in the universe that cosmologists haven’t yet accounted for.  For most scientists, this is a reason for excitement.  This kind of issue typically leads to new insights.

However, it’s led to a debate that someone has been asking me to comment on.  Bjorn Ekeberg, a philosopher of science, has focused in on this problem, along with others, to assert that cosmology has some big problems, calling into question the overall big bang cosmology.  This drew a response from cosmologist and science writer Ethan Siegel pushing back against Ekeberg’s claim and accusing him of being anti-science.  Ekeberg has responded accusing Siegel of being a “temple guard” for big bang cosmology.

Name calling aside, who’s right here?  Not being a professional physicist, or knowledgeable enough to read raw physics papers, my comments are inevitably based on what various science writers have provided.

But in considering Ekeberg’s position, it’s worth reviewing the evidence for the overall big bang model.  Physicists in the 1920s figured out that, under general relativity, the universe could not be static.  It had to be either expanding or collapsing.  If it was expanding, it was smaller yesterday than today, and smaller the day before.  Following that back led to a period in the past where everything was all bunched up together, very dense and very hot.  The physics of the early universe could be mathematically deduced and predictions made.  (This led Einstein to fudge his equations a bit by adding a cosmological constant, making the universe it predicted static.)

Then in the late 1920s, Edwin Hubble discovered that the light from every galaxy beyond a certain distance was red shifted, with the amount of red shift being proportional to the distance.  Red shift is a doppler effect that happens when something is moving away from the observer.  Hubble had discovered that the universe is indeed expanding.  (Einstein concluded that the cosmological constant was his biggest blunder.)

Still, cosmology was slow to just accept the big bang model.  (It didn’t help that Hubble’s early estimates of the age of the universe had it younger than geologist estimates of the Earth’s age.)  It continued to be debated for decades, until the discovery of the cosmic background radiation in the 1960s, which provided evidence for the calculations of the physics of the early universe.  That was enough for most cosmologists.  The big bang became settled science.

As a lay person, reading this through the translations of the experts, classic big bang cosmology seems pretty solid.  I think Ekeberg, by implying it isn’t, oversells his thesis.  But it’s worth noting that this settled version doesn’t get into what caused the big bang in the first place.

Ekeberg also has issues with the ideas of dark matter and dark energy.  My understanding of these terms is that they’re essentially place holders, labels for our ignorance.  So criticism of them as theories has always struck me as premature.

The most often touted alternative to dark matter is MOND (modified Newtonian dynamics), but no simple modification to the equations seem able to account for all the observations.  Whatever is causing the rapid rotation of galaxies and other intergalactic effects seems to require something that is present in varying densities or intensities.  Dark matter may eventually be so different from matter as we understand it that the word “matter” might not be appropriate, but until then, the term really just refers to something mysterious causing varying gravitational effects.

This seems even more true for dark energy.  The fact that, against all expectations, the expansion of the universe is actually accelerating rather than decelerating, has to be caused by something, some form of unknown energy.  (Ironically, dark energy has resurrected Einstein’s cosmological constant.)

Granted, it does seem unnerving that this results in 95% of the matter and energy in the universe being unobservable and unaccounted for.  It’s easy to take this number and other measurement issues and accuse cosmologists of not knowing what they’re doing.  Easy, but I think facile.  The widely accepted theories that we now have are grounded in observation.  Anyone is free to propose alternatives, but to be taken seriously, those alternative have to account for at least as much of the data as the current theories.

I do think one area where Siegel is overconfident is cosmic inflation.  I’ve written about the concerns on this before.  Some version of inflation might turn out to be true, but I think his stance that it’s a settled issue isn’t justified yet.  And the fact that a significant portion of physicists are starting to question inflation, including some of its earliest supporters who now say it generates more issues than it solves, should make the rest of us cautious in our stance toward it.

So, does cosmology have issues?  Of course, and Siegel admits as much.  But is the overall big bang cosmology model in crisis as Ekeberg seems to contend?  I think this is vastly overstating the issues.  But only time and the data will tell.  Of course, this controversy will likely lead to more sales for Ekeberg’s book.

What do you think?  Is the overall big bang model in trouble?  Or is this just about fine tuning the details, such as the age of the universe?  If it is in trouble, what might replace it?

Posted in Zeitgeist | Tagged , , , , , , | 14 Comments

Recommendation: The Nanotech Succession

Okay, having already recommended two of Linda Nagata’s books, Vast and Edges, I finally got around to reading the first and second book of her Nanotech Succession series.  (I haven’t read the “zeroeth” book so this recommendation doesn’t include it.)

The first book, The Bohr Maker, takes place a few centuries in the future on Earth and in the solar system.  As in the other books, mind copying is an available technology.  People can grow new bodies, send copies of themselves (called “ghosts”) to do things, and have those ghosts return and reintegrate memories with their main copy.

However, at this point in the future history, humanity is ruled by a government known as the Commonwealth, which is controlled by religious conservatives.  (Although the religion is more of a Gaia mother-Earth, nature type religion rather than an Abrahamic faith.)  Adherence to religious precepts makes the Commonwealth hostile to many types of technology, including anything too posthuman, particularly artificial intelligence that is too much like humans.

As it turns out, the main character, Nikko Jiang-Tibayan, is an artificially created human.  In order to create him, his “father” had to agree to give him a limited lifespan.  At the start of the book, Nikko’s time is almost up.  His only chance to continue existing is to acquire a forbidden nanotechnology called “the Bohr maker”.

Much of the book involves Nikko’s schemes to acquire this technology, and the complications that ensue, including entangling a desperately poor woman named Phousita, who becomes infected with the forbidden nanotech, and subsequently a focal point for powerful interests and Nikko himself.

The book has a healthy dose of action with plenty of twists and turns.   I found it the best book in the series.

In Deception Well, thousands of years have passed since the first book.  Humanity has spread to the stars.  (Albeit slowly since FTL is not possible in this universe.)  As I discussed in my write-up for Vast, humanity has also come under attack from alien robotic craft, which humans have named the Chenzeme.  The story takes place in a solar system protected from the Chenzeme by an artificial nebula filled with alien nanobots.  The system has one planet, called Deception Well.

In orbit of Deception Well is a large ring shaped Chenzeme craft, called a Swan burster, a type of craft known for attacking planets and making them uninhabitable.  The craft appears to be dead, disabled by the alien nanotech in the system.  The first colonists to Deception Well, who came to study the Swan burster, built a space elevator up from Deception Well, with a city at the midpoint along its length, called Silk.  Something killed the first colonists.  The cause is thought by the characters in the story to have been a plague from Deception Well itself, a soup of alien nanotechnology.

The second wave of colonists, arriving centuries later, are refugees from a world destroyed by the Chenzeme.  Finding the dead city, they inhabit it, but wary of the plague from Deception Well, stay away from the planet itself.

The main character, Lot, is the son of a leader named Jupiter.  Jupiter leads an invading army from another system, who attempt to conquer Silk in order to go down to Deception Well and achieve “communion” with the planet’s nanotech.  The invasion is defeated, Jupiter goes missing, and Lot grows up a prisoner in Silk, convinced that Jupiter made it to the planet and is waiting for him.

Lot inherits from Jupiter an ability to chemically influence people, an ability that essentially makes people see him as charismatic, which becomes a major issue in the book.

If you’ve read Vast, then the names Nikko and Lot will be familiar to you.  Deception Well sets up the conditions that lead to the situation at the beginning of Vast.

Obviously there’s a lot going on in this book, but I overall found it less satisfying than The Bohr Maker.  Often the narrative seemed tedious and the pacing noticeably slow, a criticism of Nagata I’ve seen from other readers, but not one I felt myself until this book.  Still, I do recommend reading it if you’re going to read the rest of the series.

As I noted above, I haven’t read the “zeroeth” book of the series, Tech Heaven, and probably never will.  Nagata herself didn’t seem to consider it essential, only republishing it belatedly after the others, and the description doesn’t particularly attract me, but it’s worth noting that it exists.

As I noted in my other reviews, Nagata is not known for her fast paced narratives.  A lot of this material is more about psychological tension and exploration of heady concepts.  She constantly blurs the boundary between technology and biology.  If that and relatively hard science fiction meet your tastes, I highly recommend the overall series.

Posted in Science Fiction | Tagged , , , | 5 Comments

The unfolding argument: why Integrated Information Theory is not scientific

There’s an interesting new paper in Consciousness and Cognition on why causal theories such as IIT (integrated information theory) or RPT (recurrent processing theory) aren’t scientific:

How can we explain consciousness? This question has become a vibrant topic of neuroscience research in recent decades. A large body of empirical results has been accumulated, and many theories have been proposed. Certain theories suggest that consciousness should be explained in terms of brain functions, such as accessing information in a global workspace, applying higher order to lower order representations, or predictive coding. These functions could be realized by a variety of patterns of brain connectivity. Other theories, such as Information Integration Theory (IIT) and Recurrent Processing Theory (RPT), identify causal structure with consciousness. For example, according to these theories, feedforward systems are never conscious, and feedback systems always are. Here, using theorems from the theory of computation, we show that causal structure theories are either false or outside the realm of science.

To be clear, the main assertion of IIT and RPT, is that it isn’t the functionality of these neural networks that make them conscious, but their specific causal structures.  According to these theories, something about those specific recurrent feedback structures generate subjective experience.

The main point the authors make is that any output that can be produced from a recurrent feedback network, can also be produced from a “unfolded” feedforward network with ϕ  (phi), the metric IIT uses to supposedly measure the amount of consciousness present, equal to zero.  In addition, any output that can be produced by a feedforward network, can also be produced by a “folded” feedback network organized to have arbitrarily high levels of ϕ.  They discuss these assertions in detail in the appendices to the paper.

Of course, the proponents of IIT and RPT can always claim that even though the output (spike trains, behavior, etc) are identical, the feedforward networks are not conscious.  But the problem is that this leaves consciousness as purely epiphenomenal, a condition that has no causal influence.  A lot of people do accept this notion of consciousness, but as the paper authors note, it leaves these theories completely outside of any scientific ability to validate or falsify them.  It makes them unscientific.

Conscious seeming behavior can, in principle, be produced by networks that these theories would say are not conscious.  This brings back our old friend (nemesis?), the philosophical zombie (the behavioral variant), along with all the problems with that concept.

Why then do we see recurrent feedback networks in the brain?  Efficiency.  If you’ve ever done any computer programming, you’ll know that program loops can save a lot of memory and code.  Recurrent feedback loops have a similar role.  They enable a lot more processing to take place than could otherwise happen with the available substrate.  Although it’s always possible to “unfold” the network, at least in principle, into a larger one and accomplish the same result.

A neural network unfolded

Image credit: François Deloche via Wikipedia

But just as in computer programming, this comes at a cost in complexity and performance.  Which is probably why the neural networks in the cerebellum are predominantly feedforward.  For what that part of the brain needs to do, speed and reliability are the priorities.  But for much of what happens in the cortex, the additional computational capacity is well worth the trade off.

But this does set up a conundrum.  It means that consciousness is more likely to be associated with the recurrent feedback regions in the brain.  Not because recurrence is equivalent to consciousness, but because cognition requires a lot of processing in a tight space.  That means ϕ could still end up being a usable measure of whether a particular brain is currently conscious, but not necessarily for telling whether other systems are conscious, a nuanced distinction that I fear the proponents of IIT will ignore.

This paper gets at a nagging issue I’ve long had with IIT.  It takes an aspect of how neural systems are organized and posits that that aspect, integration, in and of itself, somehow produces subjective experience.  How exactly?  Essentially the answer seems to be magic.  It’s more a theory aimed at explaining the ghost in the machine than the functionality.  And as I’ve indicated before, there’s no evidence, at least not yet, for any ghost, whether generated or channeled by the brain.  There’s just the brain and what it does.

Unless of course I’m missing something?

The paper does make clear that functionalist theories of consciousness, such as GWT (global workspace theory), HOTT (higher order thought theory), or PPT (predictive processing theory) are unaffected by the unfolding argument.

h/t Neuroskeptic

Posted in Zeitgeist | Tagged , , , , | 37 Comments

Michael Graziano’s attention schema theory

It’s been a while since I’ve had a chance to highlight Graziano’s attention schema theory.  This brief video is the very barest of sketches, but I think it gets the main idea across.

Those of you who’ve known me for a while might remember that I was once quite taken with this theory of consciousness.  I still think it has substantial value in understanding metacognition and top down control of attention, but I no longer see it as the whole story, seeing it as part of a capability hierarchy.

Still, the attention schema theory makes a crucial point.  What we know of our own consciousness is based on an internal model of it that our brain constructs.  Like all models, it’s simplified in a way that optimizes it for adaptive feedback, not for purposes of understanding the mind.

The problem is that this model feels privileged, to the extent that the proposition that what it shows us isn’t accurate, is simply dismissed out of hand by many people.  That our external senses aren’t necessarily accurate is relatively easy to accept, but the idea that our inner senses might have the same limitations is often fiercely resisted.

But there is a wealth of scientific research showing that introspection is unreliable.  It actually functions quite well in day to day life.  It’s only when we attempt to use it as evidence for how the mind works that we run into trouble.  Introspective data that is corroborated by other empirical data is fine, but when it’s our only source of information, caution is called for.

Graziano’s contention that conscious awareness is essentially a data model puts him in the illusionist camp.  As I’ve often said, I think the illusionists are right, although I don’t like calling phenomenal consciousness an illusion, implying that it doesn’t exist, instead currently preferring the slightly less contentious assertion that it only exists subjectively, a loose and amorphous construction from various cognitive processes.

Posted in Zeitgeist | Tagged , , , , | 28 Comments

Brain inspired hardware

The Scientist has an interesting article up reporting on the progress that’s being made in neuromorphic hardware.

But the fact that computers “think” very differently than our brains do actually gives them an advantage when it comes to tasks like number crunching, while making them decidedly primitive in other areas, such as understanding human speech or learning from experience. If scientists want to simulate a brain that can match human intelligence, let alone eclipse it, they may have to start with better building blocks—computer chips inspired by our brains.

So-called neuromorphic chips replicate the architecture of the brain—that is, they talk to each other using “neuronal spikes” akin to a neuron’s action potential. This spiking behavior allows the chips to consume very little power and remain power-efficient even when tiled together into very large-scale systems.

Traditionally, artificial neural networks have been implemented with software.  While this gets at algorithms that may resemble the ones in biological nervous systems, it does so without the advantages of the physical implementation of those systems.  Essentially it’s emulating that hardware (wetware?), which in computing has always come with a performance hit, with the magnitude of the hit usually corresponding to just how different the hardware architectures are, and modern chips and nervous systems are very different.

There’s a lot of mystique associated with neural networks.  But it’s worth remembering that a neural network is basically a crowd sourcing strategy.  Instead of having one sophisticated and high performing processor, or a few of them, like the ones in modern commercial computers, the strategy involves having large numbers, millions or billions, of relatively simple processors, the neurons.

Each neuron sums up its inputs, both positive and negative (excitation and inhibitions) and fires when a threshold is reached, providing inputs to its downstream neurons.  Synapses, the connections between neurons, grow or weaken depending on usage, changing the overall flow of information.

Of course, biological neurons are cells, which come with all the complexity associated with cellular processes.  But we shouldn’t be surprised that evolution solved its computing and communication needs with cells, since in complex life it solves everything that way.

Neuromorphic computing is moving the actual hardware closer to the structure used in nervous systems.  I’d always known about the performance advantages that might bring, but apparently a lot of the power efficiency of the brain (which operates on about 20 watts) comes down to its analog features, and neuromorphic computing, by adopting hybrid analog-digital structures, appears to be reaping many of those benefits.

The article also discusses various attempts that are underway to run simulations of the brain, although at present they’re simulating simplified versions of it.  But combined with computational neuroscience, this approach may yield theoretical insights into actual biological brains.

I’ve written before about Moore’s Law petering out, and that further progress in computing will require innovative architectural changes for us to continue seeing progress.  I find it heartening that this kind of research is happening.  Too much of the industry seems caught up in the quantum computing hype, but this line of inquiry may yield results much sooner.

Posted in Zeitgeist | Tagged , , , , , | 33 Comments

Are zombies conscious?

This is not a question about philosophical zombies.  I did a post on them a while back.  (The TL;DR is that I find that whole concept ranges from incoherent to dubious, depending on the exact version.)

This post is on the zombies we see in fiction, such as Night of the Living Dead, the Resident Evil franchise, World War Z, and a host of other movies and shows.  Last week, while watching the Game of Thrones episode featuring the epic battle with the dead, and the way the zombies searched for and pursued their victims, a question suddenly occurred to me.  Are zombies, as traditionally portrayed, conscious?  (Yes, I know I’m talking about fantasy entities here, but I’m doing so to get at intuitions.)

Let’s say first what most, if not all, of the portrayals indicate zombies are not.  They’re not the original person.  This fits with the original Haitian Vodoo concept of a zombie, a reanimated but soulless corpse.  The zombies as typically portrayed appear to have no memory of their past life, have lost the ability to communicate with language, and generally seem to be cognitively limited in a number of ways.

In the case of Game of Thrones, the zombies are controlled by the White Walkers, but there appear to only be a limited number of those Walkers, so it doesn’t seem like they’re controlling the detailed movement of every zombie.  Broadly speaking, the GoT zombies have no long term will of their own, but a lot of their detailed movements appear to be left to their discretion.

And as in a lot of other fiction, the zombies seem able to search for and pursue victims.  This indicates that they have exteroception, awareness of their environment, enough that they can navigate around in it.  They also seem able to discriminate between other zombies and living humans.  And they seem to be able to focus attention on specific people or groups.

On the other hand, your typical zombie doesn’t appear to have much of any somatic sense, any sense of touch.  Or if they do, it doesn’t appear to affect them much.  For instance, zombies seem to only minimally notice when they lose body parts.  So their interoceptive sense is either missing or stunted.

This might tempt us to conclude that the zombies have no sense of their own body.  However, being able to navigate your environment, as the zombies can clearly do at least on some level, requires being able to understand your body’s existence and its relationship to that environment.  So the zombies appear to have a only a limited sense of their own body, but a sense nonetheless.

I mentioned above that zombies don’t have memory of their past life, but they also don’t appear to have any long term memories of their current existence.  In most depictions, they do seem to have short term memory and imagination, not instantly forgetting a prey just because that prey is momentarily out of sight.  But they don’t appear to have any memory beyond the last few moments or be able to imagine anything more than a few minutes into the future.

I think it’s fair to say that zombies, while they may have some limited sense of their body, have no metacognitive self awareness, but then neither do most animals.  Although the zombies also have no self concern, no survival instinct, which everything alive seems to have in some form or another.  They do have some limited affective desires, such as desiring to eat brains, kill humans, or whatever, but those affects generally aren’t oriented toward their own preservation.

I suspect it’s this last point that really nixes any intuition we might have of them being conscious.  But what do you think?  Which aspects are necessary for us to think of a system as conscious?  Which ones, if they were in a machine, might incline us to feel like that machine was conscious?

Posted in Zeitgeist | Tagged , , , | 36 Comments

Emotions, feelings, and action programs

Sean Carroll’s latest Mindscape podcast features an interview with neuroscientist Antonio Damasio:

When we talk about the mind, we are constantly talking about consciousness and cognition. Antonio Damasio wants us to talk about our feelings. But it’s not in an effort to be more touchy-feely; Damasio, one of the world’s leading neuroscientists, believes that feelings generated by the body are a crucial part of how we achieve and maintain homeostasis, which in turn is a key driver in understanding who we are. His most recent book, The Strange Order of Things: Life, Feeling, and the Making of Cultures, is an ambitious attempt to trace the role of feelings and our biological impulses in the origin of life, the nature of consciousness, and our flourishing as social, cultural beings.

Listening to Damasio reminded me of his specific use of the word “emotion” and the definitional issues that always arise when trying to discuss emotions, feelings, and affects.  For some people, these words all mean more or less the same thing.  For others they have distinct meanings.

Damasio’s use of the word “emotion” refers not to the conscious feeling, but to the underlying automatic reaction that causes it.  Early in the evolution of central nervous systems, these automatic reactions led directly to action.  But as animals evolved distance senses such as vision, smell, and hearing, these automatic reactions became more a predisposition toward a certain action, one that could be allowed or inhibited by higher reasoning systems.

On the blog, I’ve longed referred to these early automatic reactions as “reflexes” to communicate their non-conscious or pre-conscious nature, although I know use of that specific word has its issues, mostly because I’m conflating spinal cord programs with brainstem ones.  I’ve also seen the phrase “reflex arcs” used.  Damasio, in the interview, calls them “action programs”, which seems like a pretty good name.

The problem is that using the word “emotion” to refer specifically to the action program seems prone to confusion.  The word “emotion” may have originally meant externally caused motion (e-motion), but it seems like in our society it’s become hopelessly entangled with the conscious feeling, the information signals from the action program to our higher faculties.

It’s why I often avoid the word “emotion” now.  When I do use it, it’s generally to refer to the entire stack, from the triggered action program, to the habitual allowing or inhibition of the action, to the feeling that acts as an input to our reasoning faculties, the ones that decide which reflexes or habits to allow and which to inhibit.

“Affect” seems fraught with the same difficulties.  In some cases it refers to the action program, other times to the feeling.  So I use it somewhat in the same manner as “emotion”, although to me the word “affect” has broader applicability.  It seems strange to call pain or hunger an emotion, but calling them an affect feels suitable.

Damasio’s view that emotions evolved to drive an organism to maintain its homeostasis has always made a lot of sense to me.  After all, what else is pain, hunger, or fear but impulses to motivate a creature to maintain that homeostasis, to ensure that its energy levels and other parameters remain within a range of parameters that maximize its chance of survival.

The only impulses that don’t quite seem to fit are those related to reproduction.  It doesn’t seem like reproduction, in and of itself, has much to do with homeostasis.  Indeed, given that males often have to fight for the right to mate, and the burden pregnancy puts on female bodies, it can outright threaten homeostasis in many circumstances.

Here I think we have to back up further and ask why maintaining homeostasis is desirable for an organism, why survival matters.  This brings us back to the selfish gene.  (“Selfish” here being a metaphor for the naturally selected effects of genes that preserve and propagate their pattern.)  An organism is essentially a survival and gene propagation machine.  So, selfish genes lead to homeostasis, which leads to action programs, which cause feelings, so that an animal’s reasoning ability can optimize their chances for survival.

Of course, once an animal has the ability to reason, it can figure out ways to satisfy its feelings in a manner that doesn’t necessarily accomplish the agenda of its genes.  Birth control is the obvious example.

Anyway, I like the sound of “action program”.  Although the term “reflex arc” can work too, signalling their similarities with spinal cord reflexes but also the added complexity, although the word “arc” might throw some people.  Of course, others will see the word “program” as a fighting word.

Ultimately definitions are what society makes them.  Any thoughts on these terms?  Or on alternatives?

Posted in Zeitgeist | Tagged , , , , | 23 Comments

Avengers: Endgame

I saw it this weekend.  I will say that it’s an enjoyable and entertaining movie.

But it’s something of a logical mess.  I’m not spoiling much by saying that time travel features in the story.  Early in the movie, there’s discussion about how lame movie treatments of time travel typically are.  (Back to the Future and Hot Tub Time Machine get mentioned.)  I think this was included as a wink to irony, because the movie proceeds to create paradoxes all over the place.  Some are cleaned up, because doing so is important to the plot, but many others are ignored, because it would be inconvenient to the plot.

Those aren’t the only logical inconsistencies.  We’re not talking about scientific implausibilities, which you have to just accept in these kinds of movies, but places where the movie makes a statement, and then later outright contradicts itself.  In the experience of the movie, it ends up working, because everything is happening fast, loud, and with feeling and panache.  It’s only afterward, as you dwell on what happened, that the inconsistencies become glaring.

I don’t doubt that hard core fans will be able to come up with explanations for all those inconsistencies.  I used to do that myself when I was a boy reading the actual comics that these movies are based on.  Still, it’d be nice if so many of them weren’t required.

That said, it’s a Marvel movie.  You have to be willing to suspend your disbelief if you’re going to enjoy it.  And I did leave the theater satisfied.  If you’ve followed the saga to this point, you’ll definitely want to watch it.  (Although this is definitely not the movie to introduce yourself to the Marvel universe.)  I thought most of the character resolutions were pretty satisfying.

Highly recommended for popcorn entertainment.

Posted in Science Fiction | Tagged , , , , | 18 Comments

Protecting AI welfare?

John Basl and Eric Schwitzgebel have a short article at Aeon arguing that AI (artificial intelligence) should enjoy the same protection as animals do for scientific research.  They make the point that while AI is a long way off from achieving human level intelligence, it may achieve animal level intelligence, such as the intelligence of a dog or mouse, sometime in the near future.

Animal research is subject to review by IRBs (Institutional Research Boards), committees constituted to provide oversight of research into human or animal subjects, ensuring that ethical standards are followed for such research.  Basl and Schwitzgabel are arguing for similar committees to be formed for AI research.

Eric Schwitzgebel also posted the article on his blog.  What follows is the comment, slightly amended, that I left there.

I definitely think it’s right to start thinking about how AIs might compare to animals.  The usual comparisons with humans is currently far too much of a leap. Although I’m not sure we’re anywhere near dogs and mice yet.  Do we have an AI with the spatial and navigational intelligence of a fruit fly, a bee, or a fish?  Maybe at this point mammals are still too much of a leap.

But it does seem like there is a need for a careful analysis of what a system needs in order to be a subject of moral concern.  Saying it needs to be conscious isn’t helpful, because there is currently no consensus on the definition of consciousness.  Basl and Schwitzgabel mention the capability to have joy and sorrow, which seems like a useful criteria.  Essentially, does the system have something like sentience, the ability to feel, to experience both negative and positive affects?  Suffering in particular seems extremely relevant.

But what is suffering?  The Buddhists seemed to put a lot of early thought into this, identifying desire as the main ingredient, a desire that can’t be satisfied.  My knowledge of Buddhism is limited, but my understanding is that they believe we should convince ourselves out of such desires.  But not all desires are volitional.  For instance, I don’t believe I can really stop desiring not to be injured, or the desire to be alive, and it would be extremely hard to stop caring about friends and family.

For example, if I sustain an injury, the signal from the injury conflicts with the desire for my body to be whole and functional.  I will have an intense reflexive desire to do something about it. Intellectually I might know that there’s nothing I can do but wait to heal.  During the interim, I have to continuously inhibit the reflex to do something, which takes energy. But regardless, the reflex continues to fire and continuously needs to be inhibited, using up energy and disrupting rest.  This is suffering.

But involuntary desires seem like something we have due to the way our minds evolved.  Would we build machines like this (aside from cases where we’re explicitly attempting to replicate animal cognition)?  It seems like machine desires could be satisfied in a way that primal animal desires can’t, by learning that the desire can’t be satisfied at all.  Once that’s known, it’s not productive for one part of the system to keep needling another part to resolve it.

So if a machines sustains damage, damage it can’t fix, it’s not particularly productive for the machine’s control center to continuously cycle through reflex and inhibition.  One signal that the situation can’t be resolved should quiet the reflex, at least for a time.  Although it could always resurface periodically to see if a resolution has become possible.

That’s not to say that some directives might not be judged so critical that we would put them as constant desires in the system.  A caregiver’s desire to ensure the well being of their charge seems like a possible example.  But it seems like this would be something we only used judiciously.  

Another thing to consider is that these systems won’t have a survival instinct.  (Again unless we’re explicitly attempting to replicate organic minds.) That means the inability to fulfill an involuntary and persistent desire wouldn’t have the same implications for them that they do for a living system.  In other words, being turned off or dismantled would not be a solution the system feared.

So, I think we have to be careful with setting up a new regulatory regime.  The vast majority of AI research won’t involve anything even approaching these kinds of issues.  Making all such research subject to additional oversight would be bureaucratic and unproductive.  

But if the researchers are explicitly trying to create a system that might have sentience, then the oversight might be warranted.  In addition, having guidelines on what current research shows on how pain and suffering work, similar to the ones used for animal research, would probably be a good idea.

What do you think?  Is this getting too far ahead of ourselves?  Or is it passed time something like this was implemented?

Posted in Zeitgeist | Tagged , , , , | 55 Comments

The relationship between usefulness and falsifiability

There’s an article by Matthew R. Francis in Symmetry magazine garnering a lot of attention asking whether falsifiability is a useful criteria for scientific theories.

Popper wrote in his classic book The Logic of Scientific Discovery that a theory that cannot be proven false—that is, a theory flexible enough to encompass every possible experimental outcome—is scientifically useless. He wrote that a scientific idea must contain the key to its own downfall: It must make predictions that can be tested and, if those predictions are proven false, the theory must be jettisoned.

If you think about it, Popper’s criteria is simply that for a theory, a model, to be scientific, there must be something about observable reality that is different if it is true instead of false.  There are a wide variety of conditions that could satisfy this criteria.  It’s actually pretty broad.  But apparently not broad enough for some.

But where does this falsifiability requirement leave certain areas of theoretical physics? String theory, for example, involves physics on extremely small length scales unreachable by any foreseeable experiment. Cosmic inflation, a theory that explains much about the properties of the observable universe, may itself be untestable through direct observations. Some critics believe these theories are unfalsifiable and, for that reason, are of dubious scientific value.

At the same time, many physicists align with philosophers of science who identified flaws in Popper’s model, saying falsification is most useful in identifying blatant pseudoscience (the flat-Earth hypothesis, again) but relatively unimportant for judging theories growing out of established paradigms in science.

Physicist Sabine Hossenfelder has a response up on her blog that is well worth reading.  Both Francis and Hossenfelder discuss situations in which a rigid adherence to falsifiability is problematic, although Francis allows for a broader scope for it than Hossenfelder.

My own attitude, as a layperson and skeptic, is that it matters to me whether a theory has been tested, or can be tested in some plausibly foreseeable scenario.  I’ll grant that scientists need space to work on speculative ideas, but as Hossenfelder notes, if there is never any pressure to eventually find testable predictions in those ideas, then they eventually become just metaphysical philosophy.  Not that there’s anything wrong necessarily with metaphysics, but it doesn’t enjoy the credibility of science for a reason.

Anyway, there are a couple of points the Symmetry article makes that I want to comment on.

On that note, Caltech cosmologist Sean M. Carroll argues that many very useful theories have both falsifiable and unfalsifiable predictions. Some aspects may be testable in principle, but not by any experiment or observation we can perform with existing technology. Many particle physics models fall into that category, but that doesn’t stop physicists from finding them useful. SUSY as a concept may not be falsifiable, but many specific models within the broad framework certainly are. All the evidence we have for the existence of dark matter is indirect, which won’t go away even if laboratory experiments never find dark matter particles. Physicists accept the concept of dark matter because it works.

First is the observation that well established theories often make untested predictions.  Sure they do.  But those theories are well established because of the predictions that have been tested.  And it’s well worth keeping in mind which predictions haven’t been tested yet, because it’s always possible that they reflect areas where even a well established theory might eventually have to be adjusted in the future.

But the other point is the use of the word “useful” here.  In what way are the untestable theories useful?  What about them makes them useful?  How would they be different if they weren’t useful?  Do they add value to other theories, value that makes the predictions of that other theory more accurate?  If so, then congratulations, you’ve just made the useful theory falsifiable.

Or are they “useful” in some other manner involving their aesthetics or emotional appeal?  Do they give us a feeling like we’ve explained something, plugged a hole in our knowledge of how something works, but without enhancing our ability to make predictions?  If so, then this feels like what in psychology is often called a “just so” story.

Just-so stories are generally recognized as having little or no value.  They’re just bias enforcing narratives we come up with that make us feel better about how something got to be the way it is, but not in giving us any real insights.  The danger with such narratives is that if everyone is too satisfied with them, they might actually stifle investigation into areas that still need it.

(Of course, whether a particular theory has been or can be tested is inevitably a matter of judgment.  I’ve had numerous people over the years point to a well accepted scientific theory they disliked and insisted that it either wasn’t falsifiable or that there was no evidence for it, and then insist that the evidence accepted by the vast majority of scientists wasn’t actually evidence.  Even Popper had trouble with this, initially thinking that natural selection wasn’t falsifiable, a fact that delights the creationists who know about it.)

Falsifiability, in my understanding, is simply an insistence that a scientific theory must be epistemically useful, must enhance our ability to make more accurate predictions, directly or indirectly.  If it doesn’t do that, or at least pave the way in some foreseeable manner for other theories that might do it, then the notion might have value as philosophy, but presenting it as settled science is misleading and, I think, puts the credibility of science in jeopardy.

Unless of course I’m missing something?

Posted in Zeitgeist | Tagged , , , | 88 Comments