But look more closely at the technological ambition behind Summly, provided in part by D?Aloisio, who seems to be a uniquely adept programmer, and the company?s partner, the scientific research firm SRI International. The heart of Summly today is still-crude natural language processing technology that ?reads? a news story and attempts to understand it the way a human might?and then generate a summary based on that understanding. This is not just extraction?which looks for key words as clues to topics and meanings?but an attempt at abstraction, which is harder. Abstraction uses software to first understand, then summarize.
The undisputed champion of abstraction right now is IBM?s Watson computer?the one that beat everybody on Jeopardy!, and is now working with doctors at the Cleveland Clinic. Watson was a whole room of computer processors that would cost tens of millions of dollars.
IBM guys came to Iowa State and we discussed Watson with them (I'm in electrical engineering).
I've also read several pieces about news writing and machines.
I'm just waiting for the day when machines will be making all our blog posts.
Luckily as a human, if we choose, we can make an effort to find other ways to employ ourselves and make progress. Essentially that is how progress is made is when we are freed up with time and can move on to do other things.
IBM guys came to Iowa State and we discussed Watson with them (I'm in electrical engineering).
I've also read several pieces about news writing and machines.
I'm just waiting for the day when machines will be making all our blog posts.
Luckily as a human, if we choose, we can make an effort to find other ways to employ ourselves and make progress. Essentially that is how progress is made is when we are freed up with time and can move on to do other things.
you're prolly pretty blown away by this sort of tech eh? does iowa state ee program put a heavy emphasis on math? i recall from my college days the ee program, or 1 to 2 of them, had tons and tons of math.
i maxed out at linear algebra/dif eq so i never even hit the algorithm level math courses! whew, tough for me.
Probably. There's always a little luck involved. He got massive amount of press when he created summly. That isn't to say press=success, but he had a good idea AND press. Bad idea + press won't go far.
Go watch the old steve jobs lost interview and be inspired
Our sites: Femjoy.com - Softcore Nude Art featuring 100% All natural models Joymii.com - pioneers in cinematic erotica and passion
Promote some of the most-respected, and best-converting sites in the erotica niche (as proven by affiliates who have promoted our sites and others). Sign up at http://cash.femjoy.com
Ex GF Films | Grab Dollars
Up To 80% Rev-Share | 255 Day Cookie | Legal Content | Variety of Promo Tools | CCBill Program | GF Niche
james[at]grabdollars[dot]com | ICQ::611-99-zero-zero-20
you're prolly pretty blown away by this sort of tech eh? does iowa state ee program put a heavy emphasis on math? i recall from my college days the ee program, or 1 to 2 of them, had tons and tons of math.
i maxed out at linear algebra/dif eq so i never even hit the algorithm level math courses! whew, tough for me.
Yes, all engineering is math and science. Diff EQ (last semester)
It's only getting tougher, Electrical Engineering even has some emphasis on programming. I had to take a data structures Java class and an embedded systems class (included assembly).
Engineering is tough on the body, almost all the girls in there are beat. The other thing I don't like is the entitlement many of the students feel, sure you're smart, but it's not attractive. Many of them could use some social graces.
Probably. There's always a little luck involved. He got massive amount of press when he created summly. That isn't to say press=success, but he had a good idea AND press. Bad idea + press won't go far.
Go watch the old steve jobs lost interview and be inspired
Yes, all engineering is math and science. Diff EQ (last semester)
It's only getting tougher, Electrical Engineering even has some emphasis on programming. I had to take a data structures Java class and an embedded systems class (included assembly).
Engineering is tough on the body, almost all the girls in there are beat. The other thing I don't like is the entitlement many of the students feel, sure you're smart, but it's not attractive. Many of them could use some social graces.
i hear ya, i ended up having to go with a minor in cs, i went in not realizing just how much and how freaking hard the math was going to be. fortunately, the uni had a way out for guys like me, the difference between a cs minor and major is the math component.
i did get a kick out of assembly though and the lower level electronics classes but, yeah, many socially awkward moments with very smart kids.
lol, one, a cs class, the projects were always group projects and so were more difficult, groups had 1 smart person, 1 dumb person and a few in the middle. the smart one in my group emailed the group several hours after receiving the 2 week project to tell us all he's finished the entire project and submitted it for the team!
This is depressing for the average engineer who work hard all the life. I was writing 3d software for Intel, Nvidia and few others in 1990's, and when I became lazy to learn more math every day I been re-selling some russian's algorithms later (got some guy who was doing soviet rocket control programs before, lost them when they moved to Canada). There was lots of math, but no one paid more than a few $1000's for any math and algo's anywhere, since you can't easily patent that... the round borders in an icons design of Apple is easier to patent than a math formula, apparently. Also you can not keep any code or math secret, since you can decompile whatever executables (java is even easier). And let you hide that in a server side, are you sure anyone don't copy the file out. Finally, with a few million as prize, you find lots of 17 year old kids rewriting whatever algorithm after you see one in action.
From the articles I've read, summly doesn't do much more than regurgitate the first couple of sentences an article. They licensed the 'technology' from an outside company.
From the articles I've read, summly doesn't do much more than regurgitate the first couple of sentences an article. They licensed the 'technology' from an outside company.
Some kid from scriptlance could have done something like that many many times cheaper?
From the articles I've read, summly doesn't do much more than regurgitate the first couple of sentences an article. They licensed the 'technology' from an outside company.
"Summly came to SRI International with a core concept to solve the information overload problem, which is especially challenging for mobile devices because of their limited screen size," said David Israel, Ph.D., program director in the Artificial Intelligence Center at SRI International. "Building on SRI's expertise in machine learning and natural language processing, the Summly team is creating a new type of content, providing understandable and relevant summaries tailored for mobile devices."
i've read some into trying to find out how much sri international (stanford research) participated, from the best i can tell, sri's input was/in the single digit % on this project.
that could be incorrect but it's the best i could come up with, that was earlier today, i don't have the link, i might track it down later.
maybe yahoo spent $30m on hype or maybe they spent it on the algo..........hmmmm
Despite speculation on the Internet that Summly only licensed its summarizing technology from SRI International, an independent research institute, Mr. D?Aloisio said the technology was developed by Summly and that the company owned 100% of the intellectual property behind the service.
Summly's own Website says:
SRI International, with the help of the Summly team, built the summarisation technology behind Summly. They own a small share in the company and are helping us improve the algorithm.
So let's review: Yahoo bought a startup to acquire its talent.
But that talent lives 10,000 miles away, is still finishing high school, and is not actually responsible for the startup's core technology.
This has us suspicious that they are other reasons Yahoo bought Summly.
Some possible explanations:
Summly has an impressive list of angel investors, including Brian Chesky, Mark Pincus, Matt Mullenwegg, Li-Kai Shing, and Josh Kushner. These people are influential with lots of startups that Yahoo will want to acquire some day. Maybe Yahoo has decided spending $30 million to give one of their disappointing startups a safe landing is a way to get on their good side. Certainly Yahoo has some rehabilitation to do with Silicon Valley investors and entrepreneurs.
Yahoo is getting tons of press in the mainstream media for making a teenager a millionaire. The CEO is on the Today Show this morning. $30 million is not a lot of money to spend for so much "earned media," as marketing people call mentions in the press.
This is depressing for the average engineer who work hard all the life. I was writing 3d software for Intel, Nvidia and few others in 1990's, and when I became lazy to learn more math every day I been re-selling some russian's algorithms later (got some guy who was doing soviet rocket control programs before, lost them when they moved to Canada). There was lots of math, but no one paid more than a few $1000's for any math and algo's anywhere, since you can't easily patent that... the round borders in an icons design of Apple is easier to patent than a math formula, apparently. Also you can not keep any code or math secret, since you can decompile whatever executables (java is even easier). And let you hide that in a server side, are you sure anyone don't copy the file out. Finally, with a few million as prize, you find lots of 17 year old kids rewriting whatever algorithm after you see one in action.
well, i guess the good news is that there is still hope for engineers eh, while a huge accomplishment and hat's off to the kid, he didn't write the underlying code.
in hindsight, it makes sense, at 17, i wouldn't think the kid has had enough time to take the math courses needed to write such a sophisticated algo.
i would bet though that he does have a natural ability to understand complex data structures, a huge feat itself.
Comment