Two interesting studies publicized last week:
The first studies SuperAgers, defined as persons aged 80 and above, but with memories that are as sharp as those of healthy persons decades younger [http://www.kurzweilai.net/superager-brains-yield-new-clues-to-their-remarkable-memories]. Their unusual brain signature has three common components when compared with normal persons of similar ages: a thicker region of the cortex; significantly fewer tangles (a primary marker of Alzheimer’s disease), and a whopping supply of a specific neuron - von Economo - linked to higher social intelligence. It’s thought that these von Economo neurons play a critical role in the rapid transmission of behaviorally relevant information related to social interactions.
In the other [http://medicalxpress.com/news/2015-02-young-meditation-age-related-loss-gray.html], the scientists looked specifically at the association between age and gray matter. They compared 50 people who had mediated for years and 50 who didn't. People in both groups showed a loss of gray matter as they aged. But the researchers found among those who meditated, the volume of gray matter did not decline as much as it did among those who didn't. Accumulating scientific evidence that meditation has brain-altering capabilities might ultimately allow for an effective translation from research to practice, not only in the framework of healthy aging but also pathological aging.
So the question is what effect would have group meditation? Couldn't we preserve gray matter and increase number of von Economo neurons at the same time?
Sunday, February 08, 2015
Friday, August 23, 2013
Define Google
Google updates search to show synonyms, translation
http://www.androidauthority.com/google-updates-search-258608/
Sunday, August 18, 2013
Saturday, August 17, 2013
The meaning behind words (Google)
We’re on the cusp of deep learning for the masses. You can thank Google later — Tech News and Analysis: "word2vec. Google calls it “an efficient implementation of the continuous bag-of-words and skip-gram architectures for computing vector representations of words.”"
...
"Kaggle’s Howard calls word2vec the “crown jewel” of natural language processing. “It’s the English language compressed down to a list of numbers,” he said.
Word2vec is designed to run on a system as small as a single multicore machine (Google tested its underlying techniques over days across more than 100 cores on its data center servers). Its creators have shown how it can recognize the similarities among words (e.g., the countries in Europe) as well as how they’re related to other words (e.g., countries and capitals). It’s able to decipher analogical relationships (e.g., short is to shortest as big is to biggest), word classes (e.g.,carnivore and cormorant both relate to animals) and “linguistic regularities” (e.g., “vector(‘king’) – vector(‘man’) + vector(‘woman’) is close to vector(‘queen’))."
...
"Kaggle’s Howard calls word2vec the “crown jewel” of natural language processing. “It’s the English language compressed down to a list of numbers,” he said.
Word2vec is designed to run on a system as small as a single multicore machine (Google tested its underlying techniques over days across more than 100 cores on its data center servers). Its creators have shown how it can recognize the similarities among words (e.g., the countries in Europe) as well as how they’re related to other words (e.g., countries and capitals). It’s able to decipher analogical relationships (e.g., short is to shortest as big is to biggest), word classes (e.g.,carnivore and cormorant both relate to animals) and “linguistic regularities” (e.g., “vector(‘king’) – vector(‘man’) + vector(‘woman’) is close to vector(‘queen’))."
Wednesday, August 22, 2012
Why Technology Is Key to Overcoming the Language Barrier
Why Technology Is Key to Overcoming the Language Barrier (Jonathan Lichtman, senior vice president of SAIC): "The two main approaches to machine translation are generally rule-based (RBMT) and statistical (SMT). RBMT uses manually programmed rules to translate one language to another. SMT involves the use of previously translated content to determine which words and phrases have the highest possibility of conveying the correct meaning. Both have advantages and disadvantages, but by combining the two approaches together in the same engine — a hybrid approach — language pairs can be developed faster and with less pre-translated “training” data.
That is an important leap in translation technology because it addresses two main barriers to the automation and accuracy of a language pair: not having enough translated data available and the constant evolving nature of language."
'via Blog this'
That is an important leap in translation technology because it addresses two main barriers to the automation and accuracy of a language pair: not having enough translated data available and the constant evolving nature of language."
'via Blog this'
Monday, July 16, 2012
Lernout & Hauspie's Dragon Systems Purchase: a Horror Story Still Reverberating Today - NYTimes.com
We were all surprised when even Dragon Systems (like most NLP companies at that time) gave up and decided to sell to their biggest competitor, but as this article demonstrates, it wasn't that simple...
Goldman Sachs and a Sale Gone Horribly Awry - NYTimes.com: "The deal, the $580 million sale of a highflying technology company, Dragon Systems, had just been approved by its board and congratulations were being exchanged. But even then, at that moment of celebration, there was a sense that something was amiss.
The chief executive of Dragon had received a congratulatory bottle from the investment bankers representing the acquiring company, a Belgian competitor called Lernout & Hauspie. But he hadn’t heard from Dragon’s own bankers at Goldman Sachs."
'via Blog this'
Goldman Sachs and a Sale Gone Horribly Awry - NYTimes.com: "The deal, the $580 million sale of a highflying technology company, Dragon Systems, had just been approved by its board and congratulations were being exchanged. But even then, at that moment of celebration, there was a sense that something was amiss.
The chief executive of Dragon had received a congratulatory bottle from the investment bankers representing the acquiring company, a Belgian competitor called Lernout & Hauspie. But he hadn’t heard from Dragon’s own bankers at Goldman Sachs."
'via Blog this'
Wednesday, June 27, 2012
Official google.org Blog: The Endangered Languages Project: Supporting language preservation through technology and collaboration
Official google.org Blog: The Endangered Languages Project: Supporting language preservation through technology and collaboration: " Endangered Languages Project, a website for people to find and share the most up-to-date and comprehensive information about endangered languages. Documenting the 3,000+ languages that are on the verge of extinction (about half of all languages in the world) is an important step in preserving cultural diversity, honoring the knowledge of our elders and empowering our youth. Technology can strengthen these efforts by helping people create high-quality recordings of their elders (often the last speakers of a language), connecting diaspora communities through social media and facilitating language learning."
Monday, January 16, 2012
Solving the Machine Translation Problem - Forbes
Solving the Machine Translation Problem - Forbes: "The problem being of course that machines don’t actually “get” language very well. Don’t appreciate the richness of meaning, understand the rather sly way in which what looks like almost exactly the same phrase can have a very different meaning indeed dependent upon context.
The solution is simply to not have machines doing the translation. Instead, offer free language lessons to those who want to learn another language. Their test pieces, their exercises, will be the pieces that you are being paid to translate."
'via Blog this'
The solution is simply to not have machines doing the translation. Instead, offer free language lessons to those who want to learn another language. Their test pieces, their exercises, will be the pieces that you are being paid to translate."
'via Blog this'
Monday, October 31, 2011
Google Translate It already speaks 57 languages as well as a 10-year-old. How good can it get? By Jeremy Kingsley
http://www.slate.com/articles/technology/technology/2011/10/google_translate_will_google_s_computers_understand_languages_be.single.html
"A computer that translates "natural language" is the holy grail of artificial intelligence—language being so integral to our intelligence and to our humanness that to crack it would be to achieve artificial consciousness itself. But until relatively recently, attempts at it have mostly sucked. They’ve tended to mix the words of one language with the grammar of the other, getting both wrong in the process. Mostly, this is the fault of literal translation—the kind of process that translates kindergarten as children garden. Newer methods—dominated by Google—turn the problem around: Using data, statistics, and brute force, they succeed in part by their refusal to "deconstruct" language and teach meaning to computers in the traditional way."
"A computer that translates "natural language" is the holy grail of artificial intelligence—language being so integral to our intelligence and to our humanness that to crack it would be to achieve artificial consciousness itself. But until relatively recently, attempts at it have mostly sucked. They’ve tended to mix the words of one language with the grammar of the other, getting both wrong in the process. Mostly, this is the fault of literal translation—the kind of process that translates kindergarten as children garden. Newer methods—dominated by Google—turn the problem around: Using data, statistics, and brute force, they succeed in part by their refusal to "deconstruct" language and teach meaning to computers in the traditional way."
Saturday, July 30, 2011
Doug Cutting talks about Hadoop, and open source | ITworld
Doug Cutting talks about Hadoop, and open source | ITworld: "Doug Cutting has changed the way that IT does Big Data. Hadoop, the Open Source project he started, has made it so that any company with access to a rack of commodity PCs and a reasonable amount of programming skill can do the type of large scale data analysis work that was previously done only on supercomputers. Enterprises such as Amazon, eBay, Facebook and IBM, all the way to the Federal Reserve Board of Governors are taking advantage of tremendous value offered by this Open Source hit. Hadoop is a game changer."
Tuesday, July 26, 2011
Why Harry Potter's Latest Trick Is to Speak a Syrian Dialect - WSJ.com
Why Harry Potter's Latest Trick Is to Speak a Syrian Dialect - WSJ.com: "DUBAI, United Arab Emirates—When Khulud Abu-Homos, a television producer at OSN network here, decided to dub the Harry Potter movies into Arabic for distribution in the Middle East, she faced a quandary: which Arabic?
The Arab world, it turns out, isn't one world at all. It's a collection of overlapping worlds that harbor a dizzying array of diverse people, cultures and language. The rest of the world noticed this recently in the varied ways the Arab Spring democracy movements have played out."
The Arab world, it turns out, isn't one world at all. It's a collection of overlapping worlds that harbor a dizzying array of diverse people, cultures and language. The rest of the world noticed this recently in the varied ways the Arab Spring democracy movements have played out."
Monday, July 18, 2011
From Technologist to Philosopher - Manage Your Career - The Chronicle of Higher Education
From Technologist to Philosopher - Manage Your Career - The Chronicle of Higher Education: "Getting a humanities Ph.D. is the most deterministic path you can find to becoming exceptional in the industry. It is no longer just engineers who dominate our technology leadership, because it is no longer the case that computers are so mysterious that only engineers can understand what they are capable of. There is an industrywide shift toward more 'product thinking' in leadership—leaders who understand the social and cultural contexts in which our technologies are deployed.
Products must appeal to human beings, and a rigorously cultivated humanistic sensibility is a valued asset for this challenge. That is perhaps why a technology leader of the highest status—Steve Jobs—recently credited an appreciation for the liberal arts as key to his company's tremendous success with their various i-gadgets.
It is a convenient truth: You go into the humanities to pursue your intellectual passion; and it just so happens, as a by-product, that you emerge as a desired commodity for industry. Such is the halo of human flourishing."
Products must appeal to human beings, and a rigorously cultivated humanistic sensibility is a valued asset for this challenge. That is perhaps why a technology leader of the highest status—Steve Jobs—recently credited an appreciation for the liberal arts as key to his company's tremendous success with their various i-gadgets.
It is a convenient truth: You go into the humanities to pursue your intellectual passion; and it just so happens, as a by-product, that you emerge as a desired commodity for industry. Such is the halo of human flourishing."
Thursday, May 12, 2011
Business, software and languages...
class-built-apps-fortunes-nytimes: Personal Finance News from Yahoo! Finance: "After a few false starts, he created an app that let people send points and 'kisses' to friends. It struggled until Mr. Matei, who speaks several languages, translated the app. The next day, traffic jumped fivefold. He added games, and employees, and the app became one of the most popular Facebook programs in Europe. In late 2009, he sold to Zynga for an undisclosed sum."
Tuesday, March 22, 2011
What the E.C. Won’t See – Systran Introduces MT for European Commission Corpus - GLG News
What the E.C. Won’t See – Systran Introduces MT for European Commission Corpus - GLG News: "What does this prototype mean for the market? As we have note in earlier posts, the rules-based MT engines have been adding statistical modules to offer hybrid solutions, supporting application programming interfaces, and improving their user interfaces. What Systran has done with this release is demonstrate both its new engine’s ability to be tuned to specific domains and its SaaS credentials, although the first instance won’t get the usage that it might have seen prior to December 16th, 2010. In any case, this new Systran prototype shows that there is plenty of room for competition and innovation in the MT sector."
Wednesday, July 21, 2010
Computer automatically deciphers ancient language
Computer automatically deciphers ancient language: "Regina Barzilay, an associate professor in MIT’s Computer Science and Artificial Intelligence Lab, Ben Snyder, a grad student in her lab, and the University of Southern California’s Kevin Knight took that claim personally. At the Annual Meeting of the Association for Computational Linguistics in Sweden next month, they will present a paper on a new computer system that, in a matter of hours, deciphered much of the ancient Semitic language Ugaritic. In addition to helping archeologists decipher the eight or so ancient languages that have so far resisted their efforts, the work could also help expand the number of languages that automated translation systems like Google Translate can handle."
Friday, February 05, 2010
Overview (BOLD:PNG)
Overview (BOLD:PNG): "There is a pressing need to document the world's linguistic heritage while there is still time. The consequence of language shift is that many genres -- and many whole languages -- are quickly falling out of use. Professional linguists are compiling grammars and dictionaries, but this is painstaking work, and is not keeping up with the pace of language loss. This project is addressing the problem by training university students and literacy teachers to collect and curate oral texts from indigenous languages."
Language Log - The annihilation of computational linguistics at KCL
Language Log - The annihilation of computational linguistics at KCL: "But why target linguists? Of course linguists are by no means the sole target of the various restructuring exercises. In the sciences at King's the entire division of Engineering was shut, whereas a host of other disciplines have been targeted in the humanities (including Classics, Paleography, and American Studies). Linguists at KCL are scattered across a number of depts, so are an easy target for slogans like `only units with critical mass will be retained.' There had been discussions several years back about forming a linguistics department from the dozen or so linguists spread across CS, Philosophy, Greek, German, Education etc, but the administration decided against it.
Still, it's probably true that the marginal position of the field (neither hardcore humanities, nor hardcore science, nor hardcore engineering) and the widespread lack of awareness of what linguists do or why is an important contributing factor. Mark Liberman has, with some justification, been castigating linguists for their part in this over the pages of this blog. How critical this problem is is illustrated by the current crisis, though alleviating it is of course a long term project."
Still, it's probably true that the marginal position of the field (neither hardcore humanities, nor hardcore science, nor hardcore engineering) and the widespread lack of awareness of what linguists do or why is an important contributing factor. Mark Liberman has, with some justification, been castigating linguists for their part in this over the pages of this blog. How critical this problem is is illustrated by the current crisis, though alleviating it is of course a long term project."
Thursday, August 20, 2009
Language Watch: Is Gibbs Striking the Wrong "Denote"? - Political Punch
Corpus Linguistics and Politics:
Language Watch: Is Gibbs Striking the Wrong "Denote"? - Political Punch: "But is this a word favored in Washington, DC, by officials?
Apparently not.
Using the website CapitolWords, Jeff Connor-Linton, an associate professor of Linguistics at Georgetown, plugged in the relative frequency of use of the words “denote,” “note,” “say,” and “indicate” in the Congressional Record since 2002.
“Note” has been used 436 times, “say” 2,850 times, “indicate” 88 times.
“Denote” has been used zero times.
Connor-Linton calls the absence of the use of the word “striking.”"
Language Watch: Is Gibbs Striking the Wrong "Denote"? - Political Punch: "But is this a word favored in Washington, DC, by officials?
Apparently not.
Using the website CapitolWords, Jeff Connor-Linton, an associate professor of Linguistics at Georgetown, plugged in the relative frequency of use of the words “denote,” “note,” “say,” and “indicate” in the Congressional Record since 2002.
“Note” has been used 436 times, “say” 2,850 times, “indicate” 88 times.
“Denote” has been used zero times.
Connor-Linton calls the absence of the use of the word “striking.”"
Monday, May 11, 2009
New search engine aims to come up with just the right answer
New search engine aims to come up with just the right answer: "That's the idea behind Wolfram Alpha, a new search service that could be as much of a game-changer as Wikipedia or Google. Alpha, created by renowned mathematician, author, and entrepreneur Stephen Wolfram, uses fast computers and vast statistical databases to answer questions just as a human would — a human with advanced degrees in math."
Thursday, April 02, 2009
Playing with your own MT system...
Various machine translation tools are available like Apertium (GNU license), OpenLogos which is the open source version of Logos Machine Translation System, SYSTRAN which is one of the oldest Machine Translation company and Moses (GNU General Public License).
Moses is a phrase based machine translation tool for converting one language to another language. Technical details regarding this are available on Moses website, www.statmt.org/moses. In this article, is described a short step-by-step process for converting text from one language to another, for example Hindi to English using Moses machine translation tool.