From the Economist June 11 :
A recent report from McKinsey, a management consultancy, argues that the lagging performance of the country’s school pupils, particularly its poor and minority children, has wreaked more devastation on the economy than the current recession. American children have it easier than most other children in the world, including the supposedly lazy Europeans. They have one of the shortest school years anywhere, a mere 180 days compared with an average of 195 for OECD countries and more than 200 for East Asian countries. German children spend 20 more days in school than American ones, and South Koreans over a month more. Over 12 years, a 15-day deficit means American children lose out on 180 days of school, equivalent to an entire year. American children also have one of the shortest school days, six-and-a-half hours, adding up to 32 hours a week. By contrast, the school week is 37 hours in Luxembourg, 44 in Belgium, 53 in Denmark and 60 in Sweden. On top of that, American children do only about an hour’s-worth of homework a day, a figure that stuns the Japanese and Chinese. More at http://www.economist.com/world/unitedstates/displayStory.cfm?story_id=13825184
Sunday, June 14, 2009
Saturday, March 28, 2009
Lead System Integrators
EXECUTIVE SUMMARY: In the wake of recent cost overruns, schedule slips, and performance shortfalls, there is a growing concern about using private-sector lead system integrators (LSIs) for the execution of large, complex, defense-related acquisition programs. Abandoning the LSI approach will challenge the Services, because in most cases they lack the ability to manage complex programs on their own.
DISCUSSION:
In recent years, DOD acquisitions have turned to the LSI concept in large part because they have determined that they lack the in-house, technical, and project-management expertise needed to execute large, complex acquisition programs. Because private-sector firms often have better knowledge and expertise of rapidly developing commercial technologies, LSI arrangements can promote better technical innovation and overall system optimization.
LSI proponents argue the only way a complex program with numerous component contracts can be delivered, is by contracting for a seamless single integrated program led by the LSI which effectively streamlines the downselect and procurement process. In an LSI arrangement, the federal government has a contractual relationship with the LSI prime contractor, not with any subcontractors that report to the prime contractor. This lack of transparency makes government management and oversight of an acquisition program more difficult and increases the risk of cost overruns, schedule slippage, poor product quality (e.g. lack of interoperability), and inadequate system performance.
Of course, cost overruns, schedule slips, and performance shortfalls have plagued large weapon system acquisition programs since World War II, so LSI’s maybe a scapegoat for more fundamental contributing factors to these problems including requirements creep, funding instability and immature technologies.
“We’ve relied too much on contractors to do the work of government as a result of tightening budgets, a dearth of contracting expertise in the federal government, and a loss of focus on critical governmental roles and responsibilities in the management and oversight of acquisition programs,” Coast Guard Commandant Adm. Thad Allen said last April.1 In recent years, House Armed Services Committee members have been concerned about ceding too much program management to contractors, allowing cost overruns, schedule delays and other problems to go undetected until too late. HASC Chairman Gene Taylor (D-Miss.) said last month he wants to abandon the use of a lead system integrator. 2
In a Capitol Hill hearing March 26 on the nomination of Ashton Carter to be the next Pentagon acquisition chief, SASC Chairman Sen. Carl Levin (D-Mich.), took DOD to task for failing to be able report on its services outsourcing until 2011. “That’s a real problem,” Levin said. “We have contracted out so much of the services needed that we can’t even inventory the services for years.” 3
All evidence suggests it could take up to a decade to rebuild the Navy’s roster of competent engineers and acquisition managers to a level capable of managing a program like LCS. Because DOD does not have the requisite work force, the likely outcome is some revamping of the LSI construct. 4
Notes and Links:
1 Bennett, John T. U.S. reasserts control over contractors, Despite Deepwater Takeover, Many Say Gov’t Lacks Skills To Run Programs, Defense News, April 23, 2007 http://integrator.hanscom.af.mil/2007/April/04262007/04262007-21.htm
2 Kivlan, Terry. Lawmaker lays down markers on fiscal 2010 shipbuilding budget, Congress Daily , February 5, 2009
http://www.govexec.com/dailyfed/0209/020509cdpm1.htm
3 Chavanne, Bettina,H. DoD Acquisition Work Force Insufficient, Aviation Week's , March 27, 2009 http://www.military.com/features/0,15240,187701,00.html?wh=news
4 Grasso, Valerie B. Defense Acquisition: Use of Lead System Integrators (LSIs) - Background, Oversight, Issues, and Options for Congress Congressional Research Services (RS 22631), Jan 10, 2009 http://digital.library.unt.edu/govdocs/crs/data/2008/meta-crs-10699.tkl
DISCUSSION:
In recent years, DOD acquisitions have turned to the LSI concept in large part because they have determined that they lack the in-house, technical, and project-management expertise needed to execute large, complex acquisition programs. Because private-sector firms often have better knowledge and expertise of rapidly developing commercial technologies, LSI arrangements can promote better technical innovation and overall system optimization.
LSI proponents argue the only way a complex program with numerous component contracts can be delivered, is by contracting for a seamless single integrated program led by the LSI which effectively streamlines the downselect and procurement process. In an LSI arrangement, the federal government has a contractual relationship with the LSI prime contractor, not with any subcontractors that report to the prime contractor. This lack of transparency makes government management and oversight of an acquisition program more difficult and increases the risk of cost overruns, schedule slippage, poor product quality (e.g. lack of interoperability), and inadequate system performance.
Of course, cost overruns, schedule slips, and performance shortfalls have plagued large weapon system acquisition programs since World War II, so LSI’s maybe a scapegoat for more fundamental contributing factors to these problems including requirements creep, funding instability and immature technologies.
“We’ve relied too much on contractors to do the work of government as a result of tightening budgets, a dearth of contracting expertise in the federal government, and a loss of focus on critical governmental roles and responsibilities in the management and oversight of acquisition programs,” Coast Guard Commandant Adm. Thad Allen said last April.1 In recent years, House Armed Services Committee members have been concerned about ceding too much program management to contractors, allowing cost overruns, schedule delays and other problems to go undetected until too late. HASC Chairman Gene Taylor (D-Miss.) said last month he wants to abandon the use of a lead system integrator. 2
In a Capitol Hill hearing March 26 on the nomination of Ashton Carter to be the next Pentagon acquisition chief, SASC Chairman Sen. Carl Levin (D-Mich.), took DOD to task for failing to be able report on its services outsourcing until 2011. “That’s a real problem,” Levin said. “We have contracted out so much of the services needed that we can’t even inventory the services for years.” 3
All evidence suggests it could take up to a decade to rebuild the Navy’s roster of competent engineers and acquisition managers to a level capable of managing a program like LCS. Because DOD does not have the requisite work force, the likely outcome is some revamping of the LSI construct. 4
Notes and Links:
1 Bennett, John T. U.S. reasserts control over contractors, Despite Deepwater Takeover, Many Say Gov’t Lacks Skills To Run Programs, Defense News, April 23, 2007 http://integrator.hanscom.af.mil/2007/April/04262007/04262007-21.htm
2 Kivlan, Terry. Lawmaker lays down markers on fiscal 2010 shipbuilding budget, Congress Daily , February 5, 2009
http://www.govexec.com/dailyfed/0209/020509cdpm1.htm
3 Chavanne, Bettina,H. DoD Acquisition Work Force Insufficient, Aviation Week's , March 27, 2009 http://www.military.com/features/0,15240,187701,00.html?wh=news
4 Grasso, Valerie B. Defense Acquisition: Use of Lead System Integrators (LSIs) - Background, Oversight, Issues, and Options for Congress Congressional Research Services (RS 22631), Jan 10, 2009 http://digital.library.unt.edu/govdocs/crs/data/2008/meta-crs-10699.tkl
Friday, March 27, 2009
Task Force Mountain
Check out http://www.taskforcemountain.com/ which looks to be a forward leaning approach to organizational communication - in the military.
Sunday, March 22, 2009
Tuesday, March 10, 2009
Wolfram Alpha: 'A new paradigm for using computers and the web'
Another week another Google killer. Last week, it was Twitter as Google killer. This week it’s Wolfam Alpha. The difference with Wolfram Alpha is that it has the pedigree, engineering heft and perhaps a better mousetrap to actually live up to the billing.
Techmeme is a flutter with talk of Wolfram Alpha. Dan Farber notes that Stephen Wolfram is a scientist who has recorded a few breakthroughs and a little controversy. In a nutshell, Wolfram Alpha blends natural language, a new search model and an algorithm that takes all the data on the Web and makes it “computable.” Wolfram just recently outlined his latest creation and added:
I think it’s going to be pretty exciting. A new paradigm for using computers and the web.
Dan writes about Wolfram:
He received his Ph.D. in theoretical physics from Caltech in 1979 when he was 20 and has focused most of his career on probing complex systems. In 1988 he launched Mathematica, powerful computational software that has become the gold standard in its field. In 2002, Wolfram produced a 1,280-page tome, A New Kind of Science, based on a decade of exploration in cellular automata and complex systems.
In May, Wolfram will launch Wolfram Alpha, which is dubbed a computational knowledge engine. It’s pretty clear, what Web giant Wolfram Alpha is targeting:
Look familiar?
For the brainiacs in the house, Nova Spivak has a long post outlining Wolfram Alpha (it’s a must read). Simply put, if Spivak’s outline is only half on target Wolfram Alpha could be big.
Spivak writes:
In a nutshell, Wolfram and his team have built what he calls a “computational knowledge engine” for the Web. OK, so what does that really mean? Basically it means that you can ask it factual questions and it computes answers for you.
It doesn’t simply return documents that (might) contain the answers, like Google does, and it isn’t just a giant database of knowledge, like the Wikipedia. It doesn’t simply parse natural language and then use that to retrieve documents, like Powerset, for example.
Instead, Wolfram Alpha actually computes the answers to a wide range of questions — like questions that have factual answers such as “What country is Timbuktu in?” or “How many protons are in a hydrogen atom?” or “What is the average rainfall in Seattle this month?,” “What is the 300th digit of Pi?,” “where is the ISS?” or “When was GOOG worth more than $300?”
Think about that for a minute. It computes the answers. Wolfram Alpha doesn’t simply contain huge amounts of manually entered pairs of questions and answers, nor does it search for answers in a database of facts. Instead, it understands and then computes answers to certain kinds of questions.
Spivak later mentions that Wolfram Alpha isn’t designed to be HAL 9000. That’s refreshing. Wolfram Alpha sounds impressive, but it would be premature to call it a Google killer though. In fact, if Wolfram Alpha lives up to its billing it will be acquired at some ridiculous price either by Google or some company—Microsoft—looking to kill Google.
Techmeme is a flutter with talk of Wolfram Alpha. Dan Farber notes that Stephen Wolfram is a scientist who has recorded a few breakthroughs and a little controversy. In a nutshell, Wolfram Alpha blends natural language, a new search model and an algorithm that takes all the data on the Web and makes it “computable.” Wolfram just recently outlined his latest creation and added:
I think it’s going to be pretty exciting. A new paradigm for using computers and the web.
Dan writes about Wolfram:
He received his Ph.D. in theoretical physics from Caltech in 1979 when he was 20 and has focused most of his career on probing complex systems. In 1988 he launched Mathematica, powerful computational software that has become the gold standard in its field. In 2002, Wolfram produced a 1,280-page tome, A New Kind of Science, based on a decade of exploration in cellular automata and complex systems.
In May, Wolfram will launch Wolfram Alpha, which is dubbed a computational knowledge engine. It’s pretty clear, what Web giant Wolfram Alpha is targeting:
Look familiar?
For the brainiacs in the house, Nova Spivak has a long post outlining Wolfram Alpha (it’s a must read). Simply put, if Spivak’s outline is only half on target Wolfram Alpha could be big.
Spivak writes:
In a nutshell, Wolfram and his team have built what he calls a “computational knowledge engine” for the Web. OK, so what does that really mean? Basically it means that you can ask it factual questions and it computes answers for you.
It doesn’t simply return documents that (might) contain the answers, like Google does, and it isn’t just a giant database of knowledge, like the Wikipedia. It doesn’t simply parse natural language and then use that to retrieve documents, like Powerset, for example.
Instead, Wolfram Alpha actually computes the answers to a wide range of questions — like questions that have factual answers such as “What country is Timbuktu in?” or “How many protons are in a hydrogen atom?” or “What is the average rainfall in Seattle this month?,” “What is the 300th digit of Pi?,” “where is the ISS?” or “When was GOOG worth more than $300?”
Think about that for a minute. It computes the answers. Wolfram Alpha doesn’t simply contain huge amounts of manually entered pairs of questions and answers, nor does it search for answers in a database of facts. Instead, it understands and then computes answers to certain kinds of questions.
Spivak later mentions that Wolfram Alpha isn’t designed to be HAL 9000. That’s refreshing. Wolfram Alpha sounds impressive, but it would be premature to call it a Google killer though. In fact, if Wolfram Alpha lives up to its billing it will be acquired at some ridiculous price either by Google or some company—Microsoft—looking to kill Google.
Wednesday, March 4, 2009
Recent article in KM World on the Semantic Web:
<http://www.kmworld.com/Articles/News/News-Analysis/The-dream-of-the-Semantic-Web--52764.aspx>
The Semantic Web is going to succeed, in the face of its grandest ambitions.
When Tim Berners-Lee first proposed the idea, it promised something big. Very big. In his article in Scientific American, the vision he sketched for us had the world’s data interoperating smoothly. But the vision was actually even bigger than that. It came out of the idea of knowledge representation, which in turn sprang from work in artificial intelligence. This initial dream of the Semantic Web wasn’t just about applications sharing data. Rather, the Semantic Web promised to put this data together in ways that understood the relationships to such a degree that smart programs would be able to make logical leaps to derive new information. So, it would know that if the #86 bus runs from Cleveland Circle to Harvard Square, it would also be able to deduce that the bus has a motor, that it might occasionally need oil, that people ride in it, that they perhaps pay for the privilege, that it therefore might have some form of money collecting device and that it won’t fit in your pocket.
This is because the Semantic Web doesn’t just give us a way to share data. Any ol’ standard lets us do that, whether it’s JPG for sharing images or comma-delimited files for sharing spreadsheets. The Semantic Web gives us a way to capture the relationships among the pieces. It does this through the Semantic Web’s preferred data capture standard, RDF (Resource Description Framework). RDF captures information in what philosophers used to called judgments: "A relates to B." But unlike the usual links in HTML documents, RDF lets you specify what the relationship is: A is the sister of B, B has a height of 6 feet, B is a hedge fund manager. From this, a computer could conclude that A’s sister is six feet tall. With a little more information about how relationships work, the computer could also figure out that A is a woman and that B would probably be a good person for A to ask for a loan. These sorts of relationships are specified in what the Semantic Web calls ontologies, which are themselves expressed in a standard called OWL (Web Ontology Language).
All this is great. It enables the network to be smarter, and not just in the I-know-more-facts-than-you sort of way, because ontologies express relationships as well as facts. And because ontologies are expressed in standard formats, these smarts are cumulative. Brilliant. That Tim Berners-Lee is one heck of a smart guy. (And, it cannot be said too often, his generosity in making the Web’s standards fully open is epochal. Thank you, Sir Tim!)
So, what’s the problem? There is no problem. The Semantic Web is a great idea. It’s just not the only idea. We are always going to know, understand and care about more than any knowledge representation system can keep up with. That’s not just because there is so much to know. It’s also because so much of what we know is difficult to express precisely. Ontologies express knowledge but also necessarily (in almost all cases) clean it up, which means simplifying and specifying. There’s no harm there, so long as we remember what we’re doing, but it does mean that ontologies are tools good for some types of tasks and not very good for others. It also suggests that a system that is composed of lots of small ontologies loosely joined—and multiple ontologies covering the same fields in different ways—will capture more knowledge and be more robust than single ontologies that cover huge fields. Multiple messy ontologies include more of how the world seems to multiple, messy people. The Semantic Web’s value will grow as it becomes as inconsistent, ambiguous and imperfect as our own collective knowledge is.
Of course, the Semantic Web is only part of what we need. We will always require the help of smart, opinionated, knowledgeable people who direct us to what seems interesting and important to them. They’re going to do that by posting links and explaining why those links matter. And then they’re going to argue against the very places they’re pointing us to: "You have read what this person says! It’s totally wrong!" That’s semantics, too. And we’re going to need more and more and more of it if we are to make any sense of our world.
But, how will we find those people? And how will we be sure that they’re finding everything worth our time? And how can we make sure we’re not just listening to people who agree with us? And how will we be sure that what they claim is true isn’t a passel of lies?
We can’t. No amount of knowledge representation, computerized smarts or human effort can guarantee any of that. Knowledge will always be hit or miss, whether it’s the systematized Semantic Web or the World Wide Web’s wildly unsystematic ecosystem of wise guys, savants and know-it-alls. That’s inevitable. And, when you think about it, the alternative would be intensely undesirable.
The Semantic Web is going to succeed, in the face of its grandest ambitions.
When Tim Berners-Lee first proposed the idea, it promised something big. Very big. In his article in Scientific American, the vision he sketched for us had the world’s data interoperating smoothly. But the vision was actually even bigger than that. It came out of the idea of knowledge representation, which in turn sprang from work in artificial intelligence. This initial dream of the Semantic Web wasn’t just about applications sharing data. Rather, the Semantic Web promised to put this data together in ways that understood the relationships to such a degree that smart programs would be able to make logical leaps to derive new information. So, it would know that if the #86 bus runs from Cleveland Circle to Harvard Square, it would also be able to deduce that the bus has a motor, that it might occasionally need oil, that people ride in it, that they perhaps pay for the privilege, that it therefore might have some form of money collecting device and that it won’t fit in your pocket.
This is because the Semantic Web doesn’t just give us a way to share data. Any ol’ standard lets us do that, whether it’s JPG for sharing images or comma-delimited files for sharing spreadsheets. The Semantic Web gives us a way to capture the relationships among the pieces. It does this through the Semantic Web’s preferred data capture standard, RDF (Resource Description Framework). RDF captures information in what philosophers used to called judgments: "A relates to B." But unlike the usual links in HTML documents, RDF lets you specify what the relationship is: A is the sister of B, B has a height of 6 feet, B is a hedge fund manager. From this, a computer could conclude that A’s sister is six feet tall. With a little more information about how relationships work, the computer could also figure out that A is a woman and that B would probably be a good person for A to ask for a loan. These sorts of relationships are specified in what the Semantic Web calls ontologies, which are themselves expressed in a standard called OWL (Web Ontology Language).
All this is great. It enables the network to be smarter, and not just in the I-know-more-facts-than-you sort of way, because ontologies express relationships as well as facts. And because ontologies are expressed in standard formats, these smarts are cumulative. Brilliant. That Tim Berners-Lee is one heck of a smart guy. (And, it cannot be said too often, his generosity in making the Web’s standards fully open is epochal. Thank you, Sir Tim!)
So, what’s the problem? There is no problem. The Semantic Web is a great idea. It’s just not the only idea. We are always going to know, understand and care about more than any knowledge representation system can keep up with. That’s not just because there is so much to know. It’s also because so much of what we know is difficult to express precisely. Ontologies express knowledge but also necessarily (in almost all cases) clean it up, which means simplifying and specifying. There’s no harm there, so long as we remember what we’re doing, but it does mean that ontologies are tools good for some types of tasks and not very good for others. It also suggests that a system that is composed of lots of small ontologies loosely joined—and multiple ontologies covering the same fields in different ways—will capture more knowledge and be more robust than single ontologies that cover huge fields. Multiple messy ontologies include more of how the world seems to multiple, messy people. The Semantic Web’s value will grow as it becomes as inconsistent, ambiguous and imperfect as our own collective knowledge is.
Of course, the Semantic Web is only part of what we need. We will always require the help of smart, opinionated, knowledgeable people who direct us to what seems interesting and important to them. They’re going to do that by posting links and explaining why those links matter. And then they’re going to argue against the very places they’re pointing us to: "You have read what this person says! It’s totally wrong!" That’s semantics, too. And we’re going to need more and more and more of it if we are to make any sense of our world.
But, how will we find those people? And how will we be sure that they’re finding everything worth our time? And how can we make sure we’re not just listening to people who agree with us? And how will we be sure that what they claim is true isn’t a passel of lies?
We can’t. No amount of knowledge representation, computerized smarts or human effort can guarantee any of that. Knowledge will always be hit or miss, whether it’s the systematized Semantic Web or the World Wide Web’s wildly unsystematic ecosystem of wise guys, savants and know-it-alls. That’s inevitable. And, when you think about it, the alternative would be intensely undesirable.
TR10: Intelligent Software Assistant
Here is an interesting article from MIT Technology Review
http://www.technologyreview.com/printer_friendly_article.aspx?id=22117&channel=specialsections§ion=tr10
March/April 2009
TR10: Intelligent Software Assistant
Adam Cheyer is leading the design of powerful software that acts as a personal aide.
By Erica Naone
Search is the gateway to the Internet for most people; for many of us, it has become second nature to distill a task into a set of keywords that will lead to the required tools and information. But Adam Cheyer, cofounder of Silicon Valley startup Siri, envisions a new way for people to interact with the services available on the Internet: a "do engine" rather than a search engine. Siri is working on virtual personal-assistant software, which would help users complete tasks rather than just collect information.
Cheyer, Siri's vice president of engineering, says that the software takes the user's context into account, making it highly useful and flexible. "In order to get a system that can act and reason, you need to get a system that can interact and understand," he says.
Siri traces its origins to a military-funded artificial-intelligence project called CALO, for "cognitive assistant that learns and organizes," that is based at the research institute SRI International. The project's leaders--including Cheyer--combined traditionally isolated approaches to artificial intelligence to try to create a personal-assistant program that improves by interacting with its user. Cheyer, while still at SRI, took a team of engineers aside and built a sample consumer version; colleagues finally persuaded him to start a company based on the prototype. Siri licenses its core technology from SRI.
Mindful of the sometimes spectacular failure of previous attempts to create a virtual personal assistant, Siri's founders have set their sights conservatively. The initial version, to be released this year, will be aimed at mobile users and will perform only specific types of functions, such as helping make reservations at restaurants, check flight status, or plan weekend activities. Users can type or speak commands in casual sentences, and the software deciphers their intent from the context. Siri is connected to multiple online services, so a quick interaction with it can accomplish several small tasks that would normally require visits to a number of websites. For example, a user can ask Siri to find a midpriced Chinese restaurant in a specific part of town and make a reservation there.
Recent improvements in computer processor power have been essential in bringing this level of sophistication to a consumer product, Cheyer says. Many of CALO's abilities still can't be crammed into such products. But the growing power of mobile phones and the increasing speed of networks make it possible to handle some of the processing at Siri's headquarters and pipe the results back to users, allowing the software to take on tasks that just couldn't be done before.
"Search does what search does very well, and that's not going anywhere anytime soon," says Dag Kittlaus, Siri's cofounder and CEO. "[But] we believe that in five years, everyone's going to have a virtual assistant to which they delegate a lot of the menial tasks."
While the software will be intelligent and useful, the company has no aspiration to make it seem human. "We think that we can create an incredible experience that will help you be more efficient in your life, in solving problems and the tasks that you do," Cheyer says. But Siri is always going to be just a tool, not a rival to human intelligence: "We're very practical minded."
See the 10 Emerging Technologies of 2009.
Weekend PlansSiri cofounder Tom Gruber volunteered Adam Cheyer to participate in a conversation with the software (shown above). Gruber explains the artificial-intelligence tasks behind its responses.
1. "The user can ask a broad question like this because Siri has information that gives clues about what the user intends. For example, the software might store data about the user's location, schedule, and past activities. Siri can deal with open-ended questions within specific areas, such as entertainment or travel."
2. "Siri pulls information relevant to the user's question from a variety of Web services and tools. In this case, it checks the weather, event listings, and directories of local attractions and uses machine learning to select certain options based on the user's past preferences. Siri can connect to various Web applications and then integrate the results into a single response."
3. "Siri interprets this reply in the context of the existing conversation, using it to refine the user's request."
4. "The software offers specific suggestions based on the user's personal preferences and its ability to categorize. Because Siri is task-oriented, rather than a search engine, it offers to buy tickets that the user selects."
5. "By now, the conversation has narrowed enough that all the user has to do is click on his choice."
6. "Siri compiles information about the event, such as band members, directions, and prices, and structures it in a logical way. It also handles the task of finding out what's available and getting the tickets."
Copyright Technology Review 2009.
http://www.technologyreview.com/printer_friendly_article.aspx?id=22117&channel=specialsections§ion=tr10
March/April 2009
TR10: Intelligent Software Assistant
Adam Cheyer is leading the design of powerful software that acts as a personal aide.
By Erica Naone
Search is the gateway to the Internet for most people; for many of us, it has become second nature to distill a task into a set of keywords that will lead to the required tools and information. But Adam Cheyer, cofounder of Silicon Valley startup Siri, envisions a new way for people to interact with the services available on the Internet: a "do engine" rather than a search engine. Siri is working on virtual personal-assistant software, which would help users complete tasks rather than just collect information.
Cheyer, Siri's vice president of engineering, says that the software takes the user's context into account, making it highly useful and flexible. "In order to get a system that can act and reason, you need to get a system that can interact and understand," he says.
Siri traces its origins to a military-funded artificial-intelligence project called CALO, for "cognitive assistant that learns and organizes," that is based at the research institute SRI International. The project's leaders--including Cheyer--combined traditionally isolated approaches to artificial intelligence to try to create a personal-assistant program that improves by interacting with its user. Cheyer, while still at SRI, took a team of engineers aside and built a sample consumer version; colleagues finally persuaded him to start a company based on the prototype. Siri licenses its core technology from SRI.
Mindful of the sometimes spectacular failure of previous attempts to create a virtual personal assistant, Siri's founders have set their sights conservatively. The initial version, to be released this year, will be aimed at mobile users and will perform only specific types of functions, such as helping make reservations at restaurants, check flight status, or plan weekend activities. Users can type or speak commands in casual sentences, and the software deciphers their intent from the context. Siri is connected to multiple online services, so a quick interaction with it can accomplish several small tasks that would normally require visits to a number of websites. For example, a user can ask Siri to find a midpriced Chinese restaurant in a specific part of town and make a reservation there.
Recent improvements in computer processor power have been essential in bringing this level of sophistication to a consumer product, Cheyer says. Many of CALO's abilities still can't be crammed into such products. But the growing power of mobile phones and the increasing speed of networks make it possible to handle some of the processing at Siri's headquarters and pipe the results back to users, allowing the software to take on tasks that just couldn't be done before.
"Search does what search does very well, and that's not going anywhere anytime soon," says Dag Kittlaus, Siri's cofounder and CEO. "[But] we believe that in five years, everyone's going to have a virtual assistant to which they delegate a lot of the menial tasks."
While the software will be intelligent and useful, the company has no aspiration to make it seem human. "We think that we can create an incredible experience that will help you be more efficient in your life, in solving problems and the tasks that you do," Cheyer says. But Siri is always going to be just a tool, not a rival to human intelligence: "We're very practical minded."
See the 10 Emerging Technologies of 2009.
Weekend PlansSiri cofounder Tom Gruber volunteered Adam Cheyer to participate in a conversation with the software (shown above). Gruber explains the artificial-intelligence tasks behind its responses.
1. "The user can ask a broad question like this because Siri has information that gives clues about what the user intends. For example, the software might store data about the user's location, schedule, and past activities. Siri can deal with open-ended questions within specific areas, such as entertainment or travel."
2. "Siri pulls information relevant to the user's question from a variety of Web services and tools. In this case, it checks the weather, event listings, and directories of local attractions and uses machine learning to select certain options based on the user's past preferences. Siri can connect to various Web applications and then integrate the results into a single response."
3. "Siri interprets this reply in the context of the existing conversation, using it to refine the user's request."
4. "The software offers specific suggestions based on the user's personal preferences and its ability to categorize. Because Siri is task-oriented, rather than a search engine, it offers to buy tickets that the user selects."
5. "By now, the conversation has narrowed enough that all the user has to do is click on his choice."
6. "Siri compiles information about the event, such as band members, directions, and prices, and structures it in a logical way. It also handles the task of finding out what's available and getting the tickets."
Copyright Technology Review 2009.
Subscribe to:
Posts (Atom)