[Part II of a long review. Part I is here.]
One area where I am deeply skeptical of Kurzweil is in the idea of scanning a living person's mind and reinstantiating them on the Net somewhere. While I think the principles of intelligence and consciousness are fairly robust, the instantaneous state of an individual's mind indeed seem to be ephemeral and delicate, dependent on the Brownian motion and diffusion of neurotransmitter molecules and the electronic state of individual neurons. Samuel Braunstein (a well-known quantum computing researcher) gave a casual talk a few years ago in which he estimated that teleporting a human being would require 10^32 bits of information. Even lopping off a couple of zeroes and doing just the brain, that's still eight or nine decimal orders of magnitude larger than the biggest data archives I know of. Braunstein is estimating at the atomic level, but at the molecular level might be adequate for much of the data; Kurzweil would abstract away a bunch of that data, but the scanning process would probably require that level of detail to capture the state of mind of a person, and it would have to be done in some small time slice. I'm not convinced that this will ultimately prove to be within the bounds of the laws of physics, let alone be technically practical. I certainly don't foresee that we'll have a brain scanner whose data rate is, say, 10^33 bits/second within fifty years.
But Kurzweil talks about an interesting "back door" to getting your mind on the Net that borrows from the cyberpunks. We already have retinal and cochlear implants, pacemakers, and experimental systems (both invasive and not) that transform neuronal impulses into mechanical movements. It seems that brain implants to help control seizures are on the horizon. It's certainly plausible that we will eventually develop implants with other capabilities -- the direct network (and starship piloting) links of science fiction, maybe trackers and immobilizers for violent felons, spinal cord bridges for those with damage. Maybe some sort of memory augmentation for those with a particular form of dementia? Then what's to stop an aging chess master with strong reasoning but fading memory from getting a memory implant that happens to have Modern Chess Openings in ROM? Kurzweil seems to hit a very important point that defining who remains human (and who becomes human -- ask Andrew of Isaac Asimov's "The Bicentennial Man") will become an interesting task over the next century or so. Baseball players can use contact lenses, weight machines and special diets but not steroids. As wearable computing becomes more unobtrusive and blends into implants, what is acceptable for professional go players, and for that matter what is detectable?
The creation of these technologies will take a long time, in my opinion; I very much doubt that any time during this century some human will abandon a physical body and move completely onto the Net. Instead, I foresee a century full of gradually more useful and interactive tools, some of which will begin to exhibit enough "intelligence" that they appear to think both strategically and spontaneously. Kurzweil talks about these personal virtual assistants, and how people will become more and more attached to them. I saw a remarkable example of this with Sony's Aibo robot dog. A friend bought one and brought it to the office. It had no real facial expressions, but when it tilted its head to one side and cocked an ear, everyone within sight said, "Oh, how cute!" projecting a personality onto a bit of clever programming. That phenomenon will undoubtedly accelerate as the behavior of computer systems becomes more complex; we will become emotionally attached to those that are helpful or friendly or clever or meet our biologically and culturally preconceived notions of cute. (Kurzweil doesn't really seem to consider deeply the possibility that the artificial intelligences we create will be *different* from us. If computer chess has taught us anything, it's that there's more than one way to be good at chess; what if the same turns out to be true of many other intellectual endeavors?)
Finally, but critically, Kurzweil seems to happily ignore the real world. Luddites make a cameo or two, mostly to serve as foils for explaining his technical notions more thoroughly, but society as anything other than a substrate for the growth of the technology itself seems to play no role in his thinking. We will all adopt the technology as soon as is feasible, and eventually we'll all move onto the Net (it'll be wonderful there). In particular, the developing world never appears in Spiritual. What will become of people in Bhutan and Tibet? Will this accelerate the disparity in birth rates between the developed and developing worlds, and what will that do to the world economy? Who will mine the coal that keeps the machines running? Does a world of physical isolation and idleness await us, and who will fix it when it breaks (as in E.M. Forster's "The Machine Stops")? Will this inflame ethnic, cultural, and religious tensions? What will Muslim and Catholic societies think of artificial intelligences that demand rights, let alone implants that mess around with any bodily function viewed as something core to being human? The Dalai Lama recently had some interesting speculations on the possibility that a true AI would have a soul, which must be drawn from the pre-existing, fixed pool of souls -- hmm, a grasshopper reincarnated as an AI?
In the end, even the predictions I agree with I find simplistic, and likely optimistic by a factor of two or more in time frame. I'm not sure if his unbridled optimism does more good than harm to the important ideas of AI, and the future of technology.
This is recommended reading, especially if you haven't spent any time thinking about the topics involved -- but fill your salt shaker first.
You'll find Kurzweil's timeline (created circa 1998, and continuing seamlessly from the birth of the Universe to the end of the twenty-first century) here.
Tuesday, July 18, 2006
The Age of Spiritual Machines, Part I
[Part I of a long review. Part II is here.]
Ray Kurzweil is a salesman, and a True Believer. I just finished reading his The Age of Spiritual Machines, in which he shares his faith in neural networks, evolutionary algorithms, Drexlerian nanotechnology, and Moore's Law, which leads him to conclude that a "strong" AI (a true intelligence, more than just a program capable of passing the Turing Test) will emerge around 2019 (indeed, will be runnable on a single PC), and that progress will continue to accelerate toward the point where human and machine intelligences merge on the Net before the end of the twenty-first century (an event which he calls the Singularity).
I have many problems with the book, though there are broad areas of agreement on fundamental principles -- I'm a believer in strong AI. If carbon-based matter can think, I see no reason why silicon-based matter can't think -- and no reason to believe that we can't build it, and that it will improve over time. But that's a very far cry from agreeing with the major themes, let alone details, of Kurzweil's book.
The first, and biggest, problem is his Law of Accelerating Returns. While Henry Adams was mulling this concept for human society a hundred years ago, Kurzweil goes far beyond Adams (whom he doesn't appear to cite, though maybe I missed it; in general, the footnoting in the book is good, but the prior literature including science fiction is certainly vast) and asserts that the evolution of the Universe itself has as a goal the creation of intelligence, and that evolution runs at an ever-accelerating pace, unstoppably. He treats this as some sort of vaguely-defined physical law, which I find implausible and poorly supported, at best. (Perhaps he has a more technical argument in a paper somewhere? After all, this is a pop "science" book.) He pays a bit of lip service to punctuated equilibria (misreading Gould, in my opinion) and the possibility of catastrophic societal meltdowns, but doesn't really put much stock in them. He doesn't deal with the fact that dinosaurs seemed quite comfortably in control of the planet until catastrophe befell them -- without any archeological or paleontological evidence that dinosaurs needed intelligence to maintain their dominance, or indeed that their evolution over much of their dominant period truly constituted "progress" as we would define it.
Likewise, Kurzweil extends Moore's Law to some sort of supernatural phenomenon, arguing that computational power starting with mechanical calculators and continuing through the end of the nominal VLSI-relevant Moore's Law in 20-30 years, then continuing through some ill-defined nanotech computational substrate, continues to accelerate. Not just stays on Moore's Law, but that the performance-doubling time will continue to shrink! While his twentieth-century chart is fascinating, I doubt very much that some sort of fundamental principle is in evidence, and that the rate of computation will continue to advance until we are computing with individual quarks. Kurzweil mentions S-curves and the end of exponential growth, but simply has faith that we will find some way around it -- that as each individual S-curve begins to tail off, there is another waiting in the wings to pick up the baton and run.
Kurzweil spends a few pages discussing quantum computing, and while it's not very good, it's also not terrible for a layman's understanding circa 1998. He does conclude (correctly, IMHO) that quantum computing is likely to be a special-purpose tool, rather than a true replacement for all computation.
Kurzweil has worked on voice recognition. I don't dispute that he dictated the bulk of Spiritual to a voice recognition system, but the assertion that keyboards would practically disappear by 2009 must have seemed a reach even in 1999. Likewise, it seems to me that he has substantially oversold the capabilities -- both contemporary impact and future breadth of applicability -- of neural nets and evolutionary algorithms. I have a little experience (more as a user than developer, in collaboration with another researcher) with both evolutionary and neural nets, and in my experience, they take a lot of care and feeding, and getting them to scale reasonably with the problem size is difficult; they tend to need fairly structured guidance, rather than simply turning them on and letting them go. Let me hasten to add that I'm a believer in the value of these technologies -- but they certainly are not yet some silver bullet that allows us to dispense with understanding problems ourselves before instructing a computer how to solve them for us.
Kurzweil believes that simple (ultra-)Moore's Law growth in computation will allow us to scale up these two technologies to the point where it's possible for us to just turn them loose (maybe with a dash or two of learning about the human brain's structure) and we'll get intelligent beings; we already have abstracted the neuron adequately, and only need to evolve large enough and correctly connected neural nets and the structures themselves will take over from there. While it's a beguiling scenario, my opinion is that we are likely to actually need new insights and will have to actively guide their development. Simply creating some sort of neuronal evolutionary soup leaves us in a combinatorial space beyond all comprehending in size -- waiting for a human brain to evolve in that environment is going to require eons, in my opinion.
Kurzweil takes on Roger Penrose's The Emperor's New Mind, which was already old news when he was writing but is still an influential book. I read TENM shortly after it came out, and while the details have long since faded, I was unconvinced by Penrose's arguments, which seemed to amount to the assertion that intelligence (or consciousness?) requires some non-physical phenomenon -- or at least new physics that we don't yet understand. In the end he comes to the suggestion that intelligence is derived at bottom from quantum processes. Let me stress that my IQ is probably half of Penrose's, and finding my accomplishments if stacked up next to his would require a microscope. I'm also not a consciousness researcher (and neither is Penrose). But I don't yet see any reason to invoke new physics (beyond possibly deepening our understanding of nonlinear dynamics and complexity). There is still a lot of wiggle room for well-understood physics to generate poorly-understood macroscopic phenomena.
So here, at least, I agree with Kurzweil: I'm not convinced by Penrose's anti-strong AI arguments (many of which, according to John McCarthy, were already well refuted before TENM was published). If intelligence is a property exhibited by matter, I see no particular reason to believe that we will always be unable to create matter that thinks.
On to part II
Ray Kurzweil is a salesman, and a True Believer. I just finished reading his The Age of Spiritual Machines, in which he shares his faith in neural networks, evolutionary algorithms, Drexlerian nanotechnology, and Moore's Law, which leads him to conclude that a "strong" AI (a true intelligence, more than just a program capable of passing the Turing Test) will emerge around 2019 (indeed, will be runnable on a single PC), and that progress will continue to accelerate toward the point where human and machine intelligences merge on the Net before the end of the twenty-first century (an event which he calls the Singularity).
I have many problems with the book, though there are broad areas of agreement on fundamental principles -- I'm a believer in strong AI. If carbon-based matter can think, I see no reason why silicon-based matter can't think -- and no reason to believe that we can't build it, and that it will improve over time. But that's a very far cry from agreeing with the major themes, let alone details, of Kurzweil's book.
The first, and biggest, problem is his Law of Accelerating Returns. While Henry Adams was mulling this concept for human society a hundred years ago, Kurzweil goes far beyond Adams (whom he doesn't appear to cite, though maybe I missed it; in general, the footnoting in the book is good, but the prior literature including science fiction is certainly vast) and asserts that the evolution of the Universe itself has as a goal the creation of intelligence, and that evolution runs at an ever-accelerating pace, unstoppably. He treats this as some sort of vaguely-defined physical law, which I find implausible and poorly supported, at best. (Perhaps he has a more technical argument in a paper somewhere? After all, this is a pop "science" book.) He pays a bit of lip service to punctuated equilibria (misreading Gould, in my opinion) and the possibility of catastrophic societal meltdowns, but doesn't really put much stock in them. He doesn't deal with the fact that dinosaurs seemed quite comfortably in control of the planet until catastrophe befell them -- without any archeological or paleontological evidence that dinosaurs needed intelligence to maintain their dominance, or indeed that their evolution over much of their dominant period truly constituted "progress" as we would define it.
Likewise, Kurzweil extends Moore's Law to some sort of supernatural phenomenon, arguing that computational power starting with mechanical calculators and continuing through the end of the nominal VLSI-relevant Moore's Law in 20-30 years, then continuing through some ill-defined nanotech computational substrate, continues to accelerate. Not just stays on Moore's Law, but that the performance-doubling time will continue to shrink! While his twentieth-century chart is fascinating, I doubt very much that some sort of fundamental principle is in evidence, and that the rate of computation will continue to advance until we are computing with individual quarks. Kurzweil mentions S-curves and the end of exponential growth, but simply has faith that we will find some way around it -- that as each individual S-curve begins to tail off, there is another waiting in the wings to pick up the baton and run.
Kurzweil spends a few pages discussing quantum computing, and while it's not very good, it's also not terrible for a layman's understanding circa 1998. He does conclude (correctly, IMHO) that quantum computing is likely to be a special-purpose tool, rather than a true replacement for all computation.
Kurzweil has worked on voice recognition. I don't dispute that he dictated the bulk of Spiritual to a voice recognition system, but the assertion that keyboards would practically disappear by 2009 must have seemed a reach even in 1999. Likewise, it seems to me that he has substantially oversold the capabilities -- both contemporary impact and future breadth of applicability -- of neural nets and evolutionary algorithms. I have a little experience (more as a user than developer, in collaboration with another researcher) with both evolutionary and neural nets, and in my experience, they take a lot of care and feeding, and getting them to scale reasonably with the problem size is difficult; they tend to need fairly structured guidance, rather than simply turning them on and letting them go. Let me hasten to add that I'm a believer in the value of these technologies -- but they certainly are not yet some silver bullet that allows us to dispense with understanding problems ourselves before instructing a computer how to solve them for us.
Kurzweil believes that simple (ultra-)Moore's Law growth in computation will allow us to scale up these two technologies to the point where it's possible for us to just turn them loose (maybe with a dash or two of learning about the human brain's structure) and we'll get intelligent beings; we already have abstracted the neuron adequately, and only need to evolve large enough and correctly connected neural nets and the structures themselves will take over from there. While it's a beguiling scenario, my opinion is that we are likely to actually need new insights and will have to actively guide their development. Simply creating some sort of neuronal evolutionary soup leaves us in a combinatorial space beyond all comprehending in size -- waiting for a human brain to evolve in that environment is going to require eons, in my opinion.
Kurzweil takes on Roger Penrose's The Emperor's New Mind, which was already old news when he was writing but is still an influential book. I read TENM shortly after it came out, and while the details have long since faded, I was unconvinced by Penrose's arguments, which seemed to amount to the assertion that intelligence (or consciousness?) requires some non-physical phenomenon -- or at least new physics that we don't yet understand. In the end he comes to the suggestion that intelligence is derived at bottom from quantum processes. Let me stress that my IQ is probably half of Penrose's, and finding my accomplishments if stacked up next to his would require a microscope. I'm also not a consciousness researcher (and neither is Penrose). But I don't yet see any reason to invoke new physics (beyond possibly deepening our understanding of nonlinear dynamics and complexity). There is still a lot of wiggle room for well-understood physics to generate poorly-understood macroscopic phenomena.
So here, at least, I agree with Kurzweil: I'm not convinced by Penrose's anti-strong AI arguments (many of which, according to John McCarthy, were already well refuted before TENM was published). If intelligence is a property exhibited by matter, I see no particular reason to believe that we will always be unable to create matter that thinks.
On to part II
Monday, July 17, 2006
Inflatable Space Station
This article talks about the success of an inflatable scale model of an inflatable space station. The full-scale thing, funded by a dude named Bigelow, is supposed to be taking guests in orbit as a hotel by 2015.
Friday, July 14, 2006
Now Available: My Ph.D. Thesis! "Architecture of a Quantum Multicomputer..."
My Ph.D. thesis, "Architecture of a Quantum Multicomputer Optimized for Shor's Factoring Algorithm," is now available from my publications page or from the arXiv as quant-ph/0607065. The arXiv version uses three slightly modified figures to dramatically reduce the size of the PostScript file. As a result, the pagination also changed. Otherwise, there are no differences.
Tuesday, July 11, 2006
Soroban v. Barbie
My older daughter's in second grade, and takes soroban lessons twice a week after school. That's right, abacus. People (well, Americans, anyway) look at me like I'm insane when I tell them that. "Do they teach her how to chip flint, too?" seems to be the thought running through their heads.
A toy she has mostly outgrown is her Barbie laptop computer (she's not really into pink and flowery, but the computer has some cool games, and talks). But lately she's been playing an addition game and some of the other math games. One of them is timed -- the faster you are, the more points you get. Recently, she was playing and getting frustrated with her ability to keep up -- so she grabbed her abacus! She's faster and more accurate at two- and three-digit addition with the abacus than in her head, and seemed to do better at the computer game with her abacus by her side!
She's also learning to multiply using the soroban, ahead of learning it in her actual second-grade class. (She can also ride a unicycle, a common hobby for grade-school girls here, and read and write several hundred kanji (characters Japan borrowed from China) already. But her English is almost non-existent at this point.)
The company that I worked for here in Japan in the early 1990s still kept its books on paper, and much of the arithmetic was done by clerks with abacuses (abaci? okay, sorobans). We also had rotary-dial telephones. And we built some fantastic technology that way.
I'm sure Japanese see just as many idiosyncracies when they move to the U.S...
A toy she has mostly outgrown is her Barbie laptop computer (she's not really into pink and flowery, but the computer has some cool games, and talks). But lately she's been playing an addition game and some of the other math games. One of them is timed -- the faster you are, the more points you get. Recently, she was playing and getting frustrated with her ability to keep up -- so she grabbed her abacus! She's faster and more accurate at two- and three-digit addition with the abacus than in her head, and seemed to do better at the computer game with her abacus by her side!
She's also learning to multiply using the soroban, ahead of learning it in her actual second-grade class. (She can also ride a unicycle, a common hobby for grade-school girls here, and read and write several hundred kanji (characters Japan borrowed from China) already. But her English is almost non-existent at this point.)
The company that I worked for here in Japan in the early 1990s still kept its books on paper, and much of the arithmetic was done by clerks with abacuses (abaci? okay, sorobans). We also had rotary-dial telephones. And we built some fantastic technology that way.
I'm sure Japanese see just as many idiosyncracies when they move to the U.S...
PGP CTO on Crypto
Jon Callas, who is CTO (and CSO) at PGP Corp., posted a great message to the IP (Interesting People) mailing list today on the difficulty of cracking crypto. It's brief but thorough, eye-opening, and very grounded in reality.
Saturday, July 08, 2006
Seventy Cents per Megabit per Second (or Less)
There was a tidbit in the paper yesterday that said that Japan has the lowest average price for broadband of anywhere in the world, at an average of about seventy cents per megabit per second. We are actually paying less -- I think about fifty bucks for 100 Mbps. This is the "gigabit family type", which means that the second hop is gigabit, which I think is shared among a maximum of 32 houses, if I remember right.
The article also said that South Korea has the best availability -- the largest percentage of households could get broadband if they wanted to.
The article also said that South Korea has the best availability -- the largest percentage of households could get broadband if they wanted to.
Thursday, July 06, 2006
Asian Conference on Quantum Information Science
The deadline for submissions to AQIS is July 15, so better get it done soon if you're submitting. I'm afraid I don't have anything ready at the moment...
AQIS is the successor conference to the EQIS series, which has been very successful here in Japan, with a strong program committee and excellent speakers over the three years I've been following it. This year it's in Beijing, so get your visa lined up, too.
AQIS is the successor conference to the EQIS series, which has been very successful here in Japan, with a strong program committee and excellent speakers over the three years I've been following it. This year it's in Beijing, so get your visa lined up, too.
Wednesday, July 05, 2006
Launches
Discovery is up, and so are the North Koreans (sort of). CNN says that the Taepodong, which is supposed to be capable of reaching the U.S. from North Korea, failed after forty seconds and fell in the Sea of Japan, outside of Japan's exclusive economic zone.
With the Yasukuni Shrine visits still causing friction between Japan and Korea and China, relationships in East Asia are tense at the moment. It seems possible that Koizumi will visit Yasukuni on the anniversary of Japan's surrender on August 15, which falls a few weeks before the end of his term as LDP president. Much of the jockeying to succeed him as president (and presumably prime minister) centers around the opinion of the contestants about the Yasukuni issue.
With the Yasukuni Shrine visits still causing friction between Japan and Korea and China, relationships in East Asia are tense at the moment. It seems possible that Koizumi will visit Yasukuni on the anniversary of Japan's surrender on August 15, which falls a few weeks before the end of his term as LDP president. Much of the jockeying to succeed him as president (and presumably prime minister) centers around the opinion of the contestants about the Yasukuni issue.
Sunday, July 02, 2006
From the Ministry of Irony
As long as I'm cribbing from the Sunday morning Daily Yomiuri, can't resist this tidbit: a man was trapped in a Schindler elevator in the building in Sendai (northern Japan) that houses the branch office of Schindler itself. He was only stuck for forty minutes or so, but the fire department had to pry the doors open to get him out. The building contains residences and other offices besides just the Schindler office; the man apparently was not a Schindler employee.
Teen Mobile Phone Use in Japan
The Daily Yomiuri has another blurb citing a MHLW study. According to this one, 92 percent of high schoolers, 48 percent of middle school students, and 24 percent of fifth and sixth graders have mobile phones. A prior survey in 2001 found 27 percent and 9 percent for the latter two categories, but the blurb in the paper doesn't say about high school students.
It also says that more than 30 percent of high schoolers use their keitai (mobile phone) more than two hours a day. I suspect this includes voice, email, i-mode browsing and games, all rolled together.
This report ought to be on MHLW's news page (in Japanese), but I'm not seeing it; it may be rolled into another report...
It also says that more than 30 percent of high schoolers use their keitai (mobile phone) more than two hours a day. I suspect this includes voice, email, i-mode browsing and games, all rolled together.
This report ought to be on MHLW's news page (in Japanese), but I'm not seeing it; it may be rolled into another report...
Japan Molecular (Nanotech?) Health Study
Today's Daily Yomiuri has a blurb from Kyodo News that says that the Health, Labor, and Welfare Ministry (Kouseiroudoushou, or MHLW) has started researching the safety of "molecular substances" used in IT, cosmetics, and more. I suspect this is a nanotechnology survey, but I'm not sure; I certainly can't imagine that no one has bothered to investigate the safety of chemicals used in chip making or makeup. There's no matching news release in Japanese or English on the MHLW website, maybe on Monday.
Hubble Camera Okay?
I had missed the news that the Hubble Space Telescope's Advanced Camera for Surveys went offline two weeks ago with some sort of power supply problem, but the ops staff apparently managed to fix it from the ground, by switching to a backup supply. Whew!
Friday, June 30, 2006
CO2: Train v. Car
A couple of weeks ago I was on a crowded Tokyo train, and there was an ad at the far end of the car with some interesting data. It showed a family of four travelling by car, and producing 880 liters of CO2 emissions. In contrast, travelling by train produced 92 liters.
I couldn't get close enough to the ad to read the details, and haven't seen it again. First question is what kind of assumptions they are making -- is this packed-to-the-gills subway versus stuck-in-traffic Hummer, or is it uncrowded shinkansen green car (first class) versus K car (sub-sub-compact)? My guess would be the that comparison is intended to be favorable to the train.
I'd also like to know how many kilometers, whether luggage is involved, etc. And how is the electricity to drive the train assumed to be produced? What losses are included? Does the gasoline figure include exploring, pumping, refining, transporting, and delivering the petroleum?
This comparison is worth what you're paying for it, but it's a start...anybody got pointers to a detailed analysis, ideally including shinkansen v. airplane?
I couldn't get close enough to the ad to read the details, and haven't seen it again. First question is what kind of assumptions they are making -- is this packed-to-the-gills subway versus stuck-in-traffic Hummer, or is it uncrowded shinkansen green car (first class) versus K car (sub-sub-compact)? My guess would be the that comparison is intended to be favorable to the train.
I'd also like to know how many kilometers, whether luggage is involved, etc. And how is the electricity to drive the train assumed to be produced? What losses are included? Does the gasoline figure include exploring, pumping, refining, transporting, and delivering the petroleum?
This comparison is worth what you're paying for it, but it's a start...anybody got pointers to a detailed analysis, ideally including shinkansen v. airplane?
Thursday, June 22, 2006
Now Available: Distributed Arithmetic on a Quantum Multicomputer
Our ISCA paper, "Distributed Arithmetic on a Quantum Multicomputer", and our JETC paper, "Architectural Implications of Quantum Computing Technologies", are now both available on the Aqua Project publications page.
500GHz Transistor
John Cressler's research group at Georgia Tech and IBM have created an SiGe chip that runs at 500GHz. They achieved this by cooling the device to 4.5K. But it wasn't a slow technology to start with, having a room-temperature switching speed of 350GHz.
If I did this right, Landauer's Limit says that at 350GHz of bit destructions @300K, each transistor would generate 1.5nanowatts. So, a billion-transistor chip would have a physical, fundamental limit of at least 1.5 watts, unless reversible computing is used. Still, we're clearly orders of magnitude from that limit... (I may have dropped a small constant there, but I think I'm within a factor of two.)
If I did this right, Landauer's Limit says that at 350GHz of bit destructions @300K, each transistor would generate 1.5nanowatts. So, a billion-transistor chip would have a physical, fundamental limit of at least 1.5 watts, unless reversible computing is used. Still, we're clearly orders of magnitude from that limit... (I may have dropped a small constant there, but I think I'm within a factor of two.)
Computer Architecture Letters & TC
In my list of fast turnaround architecture-related journals, I should have mentioned Computer Architecture Letters. Four-page letters, acceptance rate currently 22%, an official IEEE Computer Society journal, a breathtakingly strong editorial board, and good papers.
It's also worth mentioning, for those of you in quantum computing, that IEEE Transactions on Computers now has Todd Brun on its editorial board; I think this makes it the first journal with a significant architecture component (though not focus) to have someone strong in quantum computing on board. Todd also seems to be very conscientious about getting papers reviewed in a timely fashion.
It's also worth mentioning, for those of you in quantum computing, that IEEE Transactions on Computers now has Todd Brun on its editorial board; I think this makes it the first journal with a significant architecture component (though not focus) to have someone strong in quantum computing on board. Todd also seems to be very conscientious about getting papers reviewed in a timely fashion.
Friday, June 16, 2006
ISCA and Importance of Conferences
I'm leaving momentarily for the International Symposium on Computer Architecture, in Boston this year. Although Mark Oskin holds the honor of authoring the first ISCA quantum computing paper, I believe, in 2003, this will be the first time there is a full session on the topic. Three papers, one from Berkeley, one from a mixed group including Davis, Santa Barbara, and MIT, and ours (Keio, HP Labs, and NII), will be presented on the last day of the conference (Wednesday). A good performance by the presenters will go a long way to convincing the architecture community that there is important and interesting work to be done, and that direct involvement of architecture folks will accelerate the arrival of viable quantum computers.
For you physicists who are still learning about the CS conference circuit, the top CS conferences review and publish full papers only, and are very competitive (I think ISCA's acceptance rate this year was 14%). Journals are important, too, but in many ways less so. Of the top ten most-cited venues in 2003, eight are conferences and two are journals, according to CiteSeer. I think CiteSeer's publication delay list is very suspect, but it gives you an idea -- conferences are much shorter. Many ACM and IEEE journals and transactions average more than a year to return of first reviews, let alone final publication. Recognizing that this is a problem, many of the newer journals, such as JETC, JILP (well, at seven years, JILP might not count as "new"), and TACO are working hard to keep turnaround time for reviews down. But the dialog on open access is, I think, further advanced in physics than in CS, which is not what should have happened -- IMHO, CS should have been the leader on this topic.
Anyway, I'll try to blog from ISCA. The program this year looks exciting, though interestingly, there are no storage papers this year. Perhaps FAST and the new Transactions on Storage are getting the best storage papers these days?
For you physicists who are still learning about the CS conference circuit, the top CS conferences review and publish full papers only, and are very competitive (I think ISCA's acceptance rate this year was 14%). Journals are important, too, but in many ways less so. Of the top ten most-cited venues in 2003, eight are conferences and two are journals, according to CiteSeer. I think CiteSeer's publication delay list is very suspect, but it gives you an idea -- conferences are much shorter. Many ACM and IEEE journals and transactions average more than a year to return of first reviews, let alone final publication. Recognizing that this is a problem, many of the newer journals, such as JETC, JILP (well, at seven years, JILP might not count as "new"), and TACO are working hard to keep turnaround time for reviews down. But the dialog on open access is, I think, further advanced in physics than in CS, which is not what should have happened -- IMHO, CS should have been the leader on this topic.
Anyway, I'll try to blog from ISCA. The program this year looks exciting, though interestingly, there are no storage papers this year. Perhaps FAST and the new Transactions on Storage are getting the best storage papers these days?
Gedo no Senki
If you're a fan of Ursula K. Le Guin's Earthsea (I am) and the Studio Ghibli anime films (I am), you're probably excited by the upcoming "Gedo no Senki". Out next month, directed by Miyazaki Goro, son of the great Miyazaki Hayao.
The trailer is out on Google Video, and Le Guin's website has a synopsis. Le Guin says she hasn't seen the film and won't comment until she does, but that's such an exceedingly cautious and neutral comment that it makes me wonder if she has doubts.
The official web page requires Flash, and fails to detect that I have it installed for some reason.
The trailer is out on Google Video, and Le Guin's website has a synopsis. Le Guin says she hasn't seen the film and won't comment until she does, but that's such an exceedingly cautious and neutral comment that it makes me wonder if she has doubts.
The official web page requires Flash, and fails to detect that I have it installed for some reason.
Pharmaceuticals in Japan
A blurb in the Daily Yomiuri this morning led me to a report by the Office of Pharmaceutical Industry Research that says that 28 of the top 88 best-selling drugs in the world are not currently available in Japan.
The report is in Japanese, and there's a lot of specialized vocabulary I don't grok at a glance, but if I'm reading it right, Japan lags an average of 2.5 years behind the U.S. in approving drugs, with an average of 1,400 days to approval, compared to 500 for the U.S. (I have absolutely no idea what starts the clock on that approval, and I consider it quite possible that a difference in bureaucratic procedures means the reality is either better or worse.) Okay, wait, if I read this right, that's time from approval in the drug's home country to approval in the local country. That is, a European drug would appear on the U.S. market 500 days after appearing in Europe, and it would appear 1,400 days later in Japan than in Europe.
Of the top 95 drugs worldwide, 38 originated in the U.S., 14 in the U.K., 13 in Japan, 12 in Switzerland, 5 in France, 3 each in Germany, Sweden, and Denmark, and 1 each in Belgium, Israel, and Croatia.
In an unrelated article in the paper, a survey found that 35% of obstetrics departments in Japan have actually stopped delivering babies. More than a third of obstetricians are aged 60 or older, which is nominally retirement age here.
The report is in Japanese, and there's a lot of specialized vocabulary I don't grok at a glance, but if I'm reading it right, Japan lags an average of 2.5 years behind the U.S. in approving drugs, with an average of 1,400 days to approval, compared to 500 for the U.S. (I have absolutely no idea what starts the clock on that approval, and I consider it quite possible that a difference in bureaucratic procedures means the reality is either better or worse.) Okay, wait, if I read this right, that's time from approval in the drug's home country to approval in the local country. That is, a European drug would appear on the U.S. market 500 days after appearing in Europe, and it would appear 1,400 days later in Japan than in Europe.
Of the top 95 drugs worldwide, 38 originated in the U.S., 14 in the U.K., 13 in Japan, 12 in Switzerland, 5 in France, 3 each in Germany, Sweden, and Denmark, and 1 each in Belgium, Israel, and Croatia.
In an unrelated article in the paper, a survey found that 35% of obstetrics departments in Japan have actually stopped delivering babies. More than a third of obstetricians are aged 60 or older, which is nominally retirement age here.
Subscribe to:
Posts (Atom)