ftfp: "Different sets of one or more ads can be associated with the different levels of user expertise in a give topic. Using evidence of sophistication determined using user actions for example, ads targeted for novices, average sophistication, or experts (e.g., children, tourists, scientists) may be served and rendered."
This is simply an extension of what google already does at the page level. Instead of settling for targeting ads based on the contents of the page, google would like to tailor ads based on what the user is specifically looking at on a page. The above quote denotes the fact that they are likely to find correlations among certain demographic and age groups.
"In this example, one or more ads associated with topic 1 might originally be rendered in association with the document 1410. If a user were to follow the link 1414a, interest scores of one or more ads associated with topic 2 could be increased. In this case, upon returning back to document A 1410 from document B 1420, one or more ads associated with topic 2 might now be rendered in association with the document 1410."
The previous quote from the patent shows how google would use your recent browser history along with whatever tags they associate with a page to serve 'relevant' ads. This is similar to what I expect google to do with the doubleclick data they will be receiving shortly.
"9. The computer-implemented method of claim 1 wherein the actions of the user monitored are selected from a group of user actions consisting of (a) cursor positioning, (b) cursor dwell time, (c) document item selection, (d) user eye direction, (e) user facial expressions, (f) user expressions, and (g) express user topic interest input. "
Ummm, somebody at homeland security just wet their lips....
Aircraft have been using this combination of sensors for a while to handle attitude adjustments, however over time the sensors will accumulate minute errors that ultimately compound into larger ones. For this reason, an absolute reckoning system like GPS is always included.
This is a great step forward but does not mean current IR strategies are necessarily old news. The blend of these two systems holds the future.
The author seems to have an unfortunately negative attitude to the idea of robot swarms.
ftfa - ' A swarm of bees is frightening enough, but a swarm of robots is worse.'
What the author is missing is the idea that a swarm of bees is one of the most amazing examples of cooperation in the animal kingdom. Anybody would agree that, by itself, a single bee is a simple organism capable of only a very few rudimentary tasks. Yet a swarm of bees can build structures of a complexity humans are now only beginning to rival and are still the most efficient way to collect nectar and create honey.
With that said, the attractiveness of swarms is more in the cost efficiency of the robots that will eventual be part of these swarms. Like the now infamous storm bot which is able to utilize simple and relatively cheap computers around the world to create a very impressive super computer cluster to carry out tasks that would normally require very expensive hardware, the idea of robot swarms is to utilize the economies of scales to create affordable and simple robots to achieve complicated goals.
While flying in an V shape seems rather useless, it is a necessary step towards building networks of simple robots that will one day be responsible for everything from farming to repairing space stations. The idea of robot swarms lends nicely to the idea of self healing and adapting networks of bots, where a single device failure can be dealt with efficiently and quickly.
in practice, the more cores you put on a chip the more likely one of the cores is going to fail during the QA stage. companies have mitigated this by, for example Sun and the niagara processor, selling chips with a variety of cores. in other words, you can buy an 8-core niagara, or for a lot less, a 6-core version. both chips go through the exact same manufacturing process, but even these billion dollar fab-labs are prone to problems. therefore, it makes the most economic sense to tailor the product line to the very expensive manufacturing process and maximize production, thereby ultimately lowering individual chip costs.
ftfa - "This means that switchgrass ethanol delivers 540 percent of the energy used to produce it, compared with just roughly 25 percent more energy returned by corn-based ethanol according to the most optimistic studies."
corn-based ethanol therefore delivers roughly 125% of the energy used to produce it, not 24%
how did we arrive at the english version that's in tfa. babelfish: dutch -> english? not exactly sure how you can convey dutch grammatical errors in english... smells like trolling. am i missing something here?
i am not positive about this, but doesn't size also correlate roughly with average life span? sounds like the rate of evolution is tied to the biological clock in all of us that dictates growth rate and longevity.
a professor i work with on autonomous modular robotics was interviewed a couple years back on the future implications of his research. generally our goals are search and rescue missions and possible space missions (reconfigurable robots are just so much more space-friendly) and the majority of our work has been towards these two milestones. the journalist, however, arrived at the interview with the fantastic vision of shrinking down these robots to nanoscale sizes and continuously (about a dozen times) asked if these robots would one day be able to enter the human body for medical purposes...
as you can imagine, my professor wasn't too amused but eventually gave in after the twelfth time being asked the question with an emphatic 'sure, why not.' and what do you know, the headline that week was something along the lines of 'Scientists creating robots to enter the human body.' we still haven't heard from any doctors yet....
I didn't read through all the patents but I'm pretty sure worldlogic doesn't have anything on tegic when it comes to the cell phone industry.
just one obvious example of why these patents shouldn't have been granted in the first place..
IT is practical CS, sans the theory. what's more interesting to you? making a business run smoothly or making a parser run at record speeds?
if you picked optimization of a parser you should go into CS, otherwise go with IT. in the end though, I think (i'm a CS major) that CS leaves you with infinitely more options when you graduate. in IT you learn how to use the tools, in CS you learn how to make them, and use them if you need.
if i was to hire an IT guy I'd like to hear that they checked their email first thing in the morning. nothing worse than having to wait a day for IT support to get back.
i guess programmers don't need to know math. kind of like how doctors usually don't need to know too too much about pharmacology and chemical reactions. they can just memorize what's available and refer to a chart. so in light of the author's erudition lets stop teaching med school students what a hydrogen bond is, cause you know, who cares about the 1% of med students that actually care...
brilliant...
math and cs are the same way, most programmers today really on their computer skills as opposed to any theoretical backing but that doesn't make math any less important to the field. first person 3d shooters today are noted for innovations in storyline/gameplay but what made them possible in the first place was a revolution in 3d polygon representation using quaternions back in the day pioneered by the likes of John Carmack. We still need the math to make new things work. And new things are cool.
parallel programming has a lot of obvious benefits. but say i want to write parallel code do i take advantage of 2 cores only or should i focus on a robust solution that identifies the number of core and optimizes accordingly?
erlang is interesting because it presents a parallel platform that scales to the number of cores available. ultimately one of the strongest features of modern code is the ability to run across numerous hardware architectures and on diverse group of processors. as a result, it's usually best to follow the old mantra in CS to program for the 'worst case scenario' which manifests itself in single core processors today.
parallel programming isn't impossible, it's just not that worthwhile right now for 99% of coders
a lot of login boxes have a system of recording successive requests from an ip address to stop brute force login hacks. a similar method could be used here.
This would also be a great approach to a lot of NLP/Translation annotation tasks. Although these types of tasks generally require a robustness (knowing which answers to trust and which to ignore) that anonymity makes difficult.
I believe amazon.com has filed a patent for a solution to this problem which attributes every annotation input to a unique user id. They then claim to use the average accuracy over the history of that user for whittling away the, 2 out of 10 i think the patent says, worst answers.
i'm sure some form of quality control/check will be needed and i wonder if such a solution would infringe on this patent?
"But if a computer can't read such a CAPTCHA, how does the system know the correct answer to the puzzle? Here's how: Each new word that cannot be read correctly by OCR is given to a user in conjunction with another word for which the answer is already known. The user is then asked to read both words. If they solve the one for which the answer is known, the system assumes their answer is correct for the new one. The system then gives the new image to a number of other people to determine, with higher confidence, whether the original answer was correct."
....ok, but maybe slashdottit is more like slashdotted which would be more likely than slashdoted (although the idea of page hits and doting forming a syno...) which is also more likely to lead to slashdotting as opposed to slashdoting but then again wouldn't it just be easier to write/.it or/.'t but neither of those are very intuitive. gottit?
the last decade or so has also seen a minimal amount of change to the architectures that intel and amd (most common processors) support. BUT... as parallel computers become more omnipresent, the need for strong assembly language skills should increase as developers try to squeeze out all the power they can from these new wave of computers.
"People smarter then you or i have already thought about this." "So what?"
Just trying to point out the fact that this isn't as 'revolutionary' as/. would have us think. The philosophies he eschews aren't really new and have been proven very difficult to implement effieciently.
This isn't one of those 'doh, why didn't I think of that ideas' it's more like a 'can he really make this work' idea. There is no 'ultimate' package/toolkit out there to solve problems involving differential equations (even though differential equations are everywhere) because every situation is unique. And so, since smart people have already thought about this long and hard its worth any interested readers' time to look further into the field and state of the art. Given the amount of work and the amount of papers/conferences coming out of the field of AI and control theory in the last couple of decades, it's worth taking in HTM philosophies with a grain of salt and a healthy dose of skepticism.
Has Hawkins demonstrated any new solutions to previously unsolved problems or just simplified, more abstracted, brute force approaches to old/easy problems (that take much longer to run then the s.o.t.a.)?
The first is an alogorithm which utilizes forward and back-tracking "to find the unknown parameters of a hidden Markov model." The second is a similar algorithm used for learning 'known' causes (for reference).
I work in computational linguistics and the time an algorithm takes to run and the amount of memory it requires are serious limitations. That's why ad-hoc systems are so common.
"I don't think the human race will survive the next thousand years unless we spread into space... [as] there are too many accidents that can befall life on a single planet."
Like Hawkings says, we can't stop everything... or maybe that was Murphy... anyway, i'm all for interstellar travel as the ultimate solution to 'this' problem.
The problem of autonomous reconfiguration planning is still quite a barrier to these robots. The majority of work that's been done in the field has been generally ad-hoc approaches where specific control algorithms are applied to specific problems (ie. if the robots form a loop then how do you coordinate a rolling gait).
http://en.wikipedia.org/wiki/Self-Reconfiguring_Mo dular_Robotics
That's a link to a wikipedia page that does a reasonable job at identifying the currently most difficult problems researchers face in modular robot development.
http://modlab.seas.upenn.edu/
This is a link to the lab I work in, we also develop a variant of the superbot currently in the process of being refined. Click 'CKbot', then click 'movies' to see some recent clips of our bots in action, including a much improved rolling gait to the one linked to in the article.
Slashdot Mirror
ftfp: "Different sets of one or more ads can be associated with the different levels of user expertise in a give topic. Using evidence of sophistication determined using user actions for example, ads targeted for novices, average sophistication, or experts (e.g., children, tourists, scientists) may be served and rendered."
This is simply an extension of what google already does at the page level. Instead of settling for targeting ads based on the contents of the page, google would like to tailor ads based on what the user is specifically looking at on a page. The above quote denotes the fact that they are likely to find correlations among certain demographic and age groups.
"In this example, one or more ads associated with topic 1 might originally be rendered in association with the document 1410. If a user were to follow the link 1414a, interest scores of one or more ads associated with topic 2 could be increased. In this case, upon returning back to document A 1410 from document B 1420, one or more ads associated with topic 2 might now be rendered in association with the document 1410."
The previous quote from the patent shows how google would use your recent browser history along with whatever tags they associate with a page to serve 'relevant' ads. This is similar to what I expect google to do with the doubleclick data they will be receiving shortly.
On a more ominous note, the following claim is a bit unsettling and reminds me of http://yro.slashdot.org/article.pl?sid=08/03/21/1511240. Who's letting all these guys control cameras in our houses?
"9. The computer-implemented method of claim 1 wherein the actions of the user monitored are selected from a group of user actions consisting of (a) cursor positioning, (b) cursor dwell time, (c) document item selection, (d) user eye direction, (e) user facial expressions, (f) user expressions, and (g) express user topic interest input. "
Ummm, somebody at homeland security just wet their lips....
Aircraft have been using this combination of sensors for a while to handle attitude adjustments, however over time the sensors will accumulate minute errors that ultimately compound into larger ones. For this reason, an absolute reckoning system like GPS is always included.
This is a great step forward but does not mean current IR strategies are necessarily old news. The blend of these two systems holds the future.
The author seems to have an unfortunately negative attitude to the idea of robot swarms.
ftfa - ' A swarm of bees is frightening enough, but a swarm of robots is worse.'
What the author is missing is the idea that a swarm of bees is one of the most amazing examples of cooperation in the animal kingdom. Anybody would agree that, by itself, a single bee is a simple organism capable of only a very few rudimentary tasks. Yet a swarm of bees can build structures of a complexity humans are now only beginning to rival and are still the most efficient way to collect nectar and create honey.
With that said, the attractiveness of swarms is more in the cost efficiency of the robots that will eventual be part of these swarms. Like the now infamous storm bot which is able to utilize simple and relatively cheap computers around the world to create a very impressive super computer cluster to carry out tasks that would normally require very expensive hardware, the idea of robot swarms is to utilize the economies of scales to create affordable and simple robots to achieve complicated goals.
While flying in an V shape seems rather useless, it is a necessary step towards building networks of simple robots that will one day be responsible for everything from farming to repairing space stations. The idea of robot swarms lends nicely to the idea of self healing and adapting networks of bots, where a single device failure can be dealt with efficiently and quickly.
in practice, the more cores you put on a chip the more likely one of the cores is going to fail during the QA stage. companies have mitigated this by, for example Sun and the niagara processor, selling chips with a variety of cores. in other words, you can buy an 8-core niagara, or for a lot less, a 6-core version. both chips go through the exact same manufacturing process, but even these billion dollar fab-labs are prone to problems. therefore, it makes the most economic sense to tailor the product line to the very expensive manufacturing process and maximize production, thereby ultimately lowering individual chip costs.
ftfa - "This means that switchgrass ethanol delivers 540 percent of the energy used to produce it, compared with just roughly 25 percent more energy returned by corn-based ethanol according to the most optimistic studies."
corn-based ethanol therefore delivers roughly 125% of the energy used to produce it, not 24%
how did we arrive at the english version that's in tfa. babelfish: dutch -> english? not exactly sure how you can convey dutch grammatical errors in english... smells like trolling. am i missing something here?
i am not positive about this, but doesn't size also correlate roughly with average life span? sounds like the rate of evolution is tied to the biological clock in all of us that dictates growth rate and longevity.
ftfa..
Korea's #1 Starcraft Player: I would like to have a good car and a fancy girlfriend.
um.... isn't it supposed to be the other way around?
a professor i work with on autonomous modular robotics was interviewed a couple years back on the future implications of his research. generally our goals are search and rescue missions and possible space missions (reconfigurable robots are just so much more space-friendly) and the majority of our work has been towards these two milestones. the journalist, however, arrived at the interview with the fantastic vision of shrinking down these robots to nanoscale sizes and continuously (about a dozen times) asked if these robots would one day be able to enter the human body for medical purposes...
as you can imagine, my professor wasn't too amused but eventually gave in after the twelfth time being asked the question with an emphatic 'sure, why not.' and what do you know, the headline that week was something along the lines of 'Scientists creating robots to enter the human body.' we still haven't heard from any doctors yet....
Tegic, the owners of t9 started filing patents in 1996, http://www.tegic.com/about/patent-leadership.asp
t ahtml%2Fsrchnum.htm&Sect1=PTO1&Sect2=HITOFF&p=1&r= 1&l=50&f=G&d=PALL&s1=5187480.PN.&OS=PN/5187480&RS= PN/5187480
I didn't read through all the patents but I'm pretty sure worldlogic doesn't have anything on tegic when it comes to the cell phone industry.
just one obvious example of why these patents shouldn't have been granted in the first place..
here's the original t9 patent for reference: http://patft.uspto.gov/netacgi/nph-Parser?u=%2Fne
if you picked optimization of a parser you should go into CS, otherwise go with IT. in the end though, I think (i'm a CS major) that CS leaves you with infinitely more options when you graduate. in IT you learn how to use the tools, in CS you learn how to make them, and use them if you need.
if i was to hire an IT guy I'd like to hear that they checked their email first thing in the morning. nothing worse than having to wait a day for IT support to get back.
brilliant...
math and cs are the same way, most programmers today really on their computer skills as opposed to any theoretical backing but that doesn't make math any less important to the field. first person 3d shooters today are noted for innovations in storyline/gameplay but what made them possible in the first place was a revolution in 3d polygon representation using quaternions back in the day pioneered by the likes of John Carmack. We still need the math to make new things work. And new things are cool.
erlang is interesting because it presents a parallel platform that scales to the number of cores available. ultimately one of the strongest features of modern code is the ability to run across numerous hardware architectures and on diverse group of processors. as a result, it's usually best to follow the old mantra in CS to program for the 'worst case scenario' which manifests itself in single core processors today.
parallel programming isn't impossible, it's just not that worthwhile right now for 99% of coders
a lot of login boxes have a system of recording successive requests from an ip address to stop brute force login hacks. a similar method could be used here.
they could offer a button that states : "this is not text" and then the user would be given a whole new captcha
http://yro.slashdot.org/article.pl?sid=07/04/03/22 11258
I believe amazon.com has filed a patent for a solution to this problem which attributes every annotation input to a unique user id. They then claim to use the average accuracy over the history of that user for whittling away the, 2 out of 10 i think the patent says, worst answers.
i'm sure some form of quality control/check will be needed and i wonder if such a solution would infringe on this patent?
"But if a computer can't read such a CAPTCHA, how does the system know the correct answer to the puzzle? Here's how: Each new word that cannot be read correctly by OCR is given to a user in conjunction with another word for which the answer is already known. The user is then asked to read both words. If they solve the one for which the answer is known, the system assumes their answer is correct for the new one. The system then gives the new image to a number of other people to determine, with higher confidence, whether the original answer was correct."
....ok, but maybe slashdottit is more like slashdotted which would be more likely than slashdoted (although the idea of page hits and doting forming a syno...) which is also more likely to lead to slashdotting as opposed to slashdoting but then again wouldn't it just be easier to write /.it or /.'t but neither of those are very intuitive. gottit?
the last decade or so has also seen a minimal amount of change to the architectures that intel and amd (most common processors) support. BUT... as parallel computers become more omnipresent, the need for strong assembly language skills should increase as developers try to squeeze out all the power they can from these new wave of computers.
"So what?"
Just trying to point out the fact that this isn't as 'revolutionary' as
This isn't one of those 'doh, why didn't I think of that ideas' it's more like a 'can he really make this work' idea. There is no 'ultimate' package/toolkit out there to solve problems involving differential equations (even though differential equations are everywhere) because every situation is unique. And so, since smart people have already thought about this long and hard its worth any interested readers' time to look further into the field and state of the art. Given the amount of work and the amount of papers/conferences coming out of the field of AI and control theory in the last couple of decades, it's worth taking in HTM philosophies with a grain of salt and a healthy dose of skepticism.
Has Hawkins demonstrated any new solutions to previously unsolved problems or just simplified, more abstracted, brute force approaches to old/easy problems (that take much longer to run then the s.o.t.a.)?
http://en.wikipedia.org/wiki/Baum-Welch_algorithm http://en.wikipedia.org/wiki/Viterbi_algorithm
The first is an alogorithm which utilizes forward and back-tracking "to find the unknown parameters of a hidden Markov model." The second is a similar algorithm used for learning 'known' causes (for reference).
I work in computational linguistics and the time an algorithm takes to run and the amount of memory it requires are serious limitations. That's why ad-hoc systems are so common.
"I don't think the human race will survive the next thousand years unless we spread into space ... [as] there are too many accidents that can befall life on a single planet."
Like Hawkings says, we can't stop everything... or maybe that was Murphy... anyway, i'm all for interstellar travel as the ultimate solution to 'this' problem.
Question though: What language might have been used to program this thing? objective C and C
The problem of autonomous reconfiguration planning is still quite a barrier to these robots. The majority of work that's been done in the field has been generally ad-hoc approaches where specific control algorithms are applied to specific problems (ie. if the robots form a loop then how do you coordinate a rolling gait). http://en.wikipedia.org/wiki/Self-Reconfiguring_Mo dular_Robotics
That's a link to a wikipedia page that does a reasonable job at identifying the currently most difficult problems researchers face in modular robot development.
http://modlab.seas.upenn.edu/
This is a link to the lab I work in, we also develop a variant of the superbot currently in the process of being refined. Click 'CKbot', then click 'movies' to see some recent clips of our bots in action, including a much improved rolling gait to the one linked to in the article.