Songbirds use grammar rules

August 11th, 2011

Researchers have found that songbirds have something that resembles grammar as we know and are very responsive to rule violation. The birds have a syntax in their tweets, maybe not the same concepts like us (that is nouns, pronouns, verbs, adjectives, adverbs and so on), but they have a syntactic structure. Syntax is the study of principles and rules for constructing sentences and grammar rules are a part of syntax.

Language is made up of signs, meanings and a code connecting signs with their meanings. Semiotics is the study that looks at how signs and meanings are combined, used and interpreted ( you can read it up in our paper Semiotics and Intrinsic Semantics).

The research findings are published in Nature and NewScientist.

Machine Learning Startup Acquired by ai-one

August 4th, 2011

Press Release

For Immediate Release:  August 4, 2011

San Diego artificial intelligence startup acquired by leading provider of machine learning SDKs as market for advanced applications gets hot.

San Diego CA – ai-one announced today that it acquired Auto-Semantics, a local start-up providing artificial intelligence services to corporate IT departments. The acquisition is the latest in a series of joint-ventures and acquisitions by ai-one that consolidates its leadership position within the emerging market for machine learning technologies.

In less than one year from its founding, Auto-Semantics built a solid pipeline of commercial accounts to apply computational semantics to solve “big data” and marketing problems – such as modeling consumer and investor behaviors. Computational semantics is a set of technologies that enables machines to understand human language. Olin Hyde started the company to build a “smart, personalized mobile GroupOn” application to deliver coupons but was unable attract capital for that concept. “Not getting funding was a blessing – it forced me to pivot from building speculative products into providing professional services. We solved corporate IT problems using machine learning technologies. That led me to ai-one – they had the only SDK to build semantic applications.” said Hyde.

ai-one provides programming tools that enable software developers to build machine learning into applications, websites and mobile phones. The company was founded in Zurich, Switzerland in 2003 and reincorporated as a US corporation in 2009 in anticipation of going public in 2012. Unlike most technology startups, ai-one spent more than eight years developing the core technology using funding from private investors from Europe.

The technology is distributed through consulting partnerships that use it to build custom solutions for corporations and government agencies. “A lot of people ask us about who uses our technology,” said ai-one’s President Tom Marsh, “and the fact is our customers are working on very proprietary, often secret, solutions.  There are many new applications in the pipeline coming from our partners that will hit the market in late 2011 and early 2012. It’s an exciting time for us.”

Hyde met Marsh at a local MeetUp group and later, ai-one founders Walt Diggelmann and Manfred Hoffleisch at the SDSIC SuperMath conference. Auto-Semantics signed up as an ai-one Consulting Partner on the first day the program became available. “It was clear from the beginning that Olin had the pulse on what corporate CIOs were thinking,” said Marsh, “He gets how to communicate the value of our big idea: Machines can learn just like we do, and you don’t have to be IBM or Google to play in this space.  Olin fits our entrepreneurial culture with international business experience.” The stock transaction ties Hyde to ai-one where he will serve as Vice President of Business Development.

ai-one also acquired Berlin-based PPM Data Management GmbH last year and formed two joint-ventures earlier this year that embed ai-one’s advanced pattern recognition technology in commercial services: ai-ibiomics gmbh provides personalized medicine using genetic sequencing and Forensity AG sells shoeprint recognition software to law enforcement agencies and crime laboratories.

About ai-one inc., ai-one provides technologies that enable programmers to build artificial intelligence into software programs. Based in San Diego with offices in Zurich and Berlin, ai-one’s “biologically inspired intelligence” is a virtual brain that learns without human intervention. Technically described as an adaptive holosemantic data space with semiotic capabilities, ai-one’s approach provides more accurate answers than competing technologies.  The Topic-Mapper™ SDK for text enables developers to create intelligent applications that deliver better sense-making capabilities for semantic discovery, lightweight ontologies, knowledge collaboration, sentiment analysis, artificial intelligence and data mining.

Contact: Tom Marsh, Phone: 1-858-531-0674, email: tm@ai-one.com, web: www.ai-one.com

 

###

 

Search needs a shake-up

August 4th, 2011

Oren Etzioni, computer scientist at the University of Washington has penned Search needs a shake-up, the August commentary in Nature (the full commentary is available to Natures subscribers or you can get the UW’s news). The piece, in short, is a call to academics and industry researchers to revolutionise how we find information on the web.

Now, searching on the web is typing a keyword (which is just a string of characters) into the search box and the search engine goes off and searches for exactly that string of characters and presents the website with that word. Instead, Etzioni proposes that the web search engine would identify basic entities – persons, places, objects – and point out the relationships between them. Which is exactly our approach. Only, we think, this can be applied in all kinds of documents and not only on the web. Our technology detects intrinsic semantic structure of the text data and works with lightweight ontologies that show the associations and significance of every element. Watch our webcast for more info.

Overview Video of ai-one Topic-Mapper SDK for Machine Learning Applications

August 1st, 2011

Click this link for an video overview of the ai-one Topic-Mapper SDK for machine learning applications (8 minutes).

Browser extension Hyperwords

July 27th, 2011

The browser plugin Hyperwords is based on the research of Doug Engelhbart and turns words and numbers into hyperlinks. Lets first take a look at how to use Hyperwords.

 

 

 

Hyperwords not only lets you jump to another page or another web site (like “normal” hyperlinks). Instead Hyperwords lets you interact with the word in several ways. After selecting text, a small blue ball and then a pop-up menu appear, offering reference, sharing and (currency) conversion, even translation options. These options can be customised and expanded to your taste.

Hyperwords allows us to set in context what we read, associate the words (and numbers) just like in an associative network (See Prof. Dr. Ulrich Reimer explanation of Lightweight Ontologies LWO).

You can download and install Hyperwords on Firefox, Chrome or Safari. Head over to hyperwords.net.

Computer reads manuals and learns how to take over the (virtual) world

July 20th, 2011

And not only that, it also learns language. Researchers have designed a computer system that can understand language, that is the instructions to play Civilisation II. The video games industry surely can’t wait to get their hands on this. Imagine this: Really smart game opponents.. (via MIT news).

Machine Learning Makes Twitter Smarter in Portuguese

July 14th, 2011

Press Release

Brazilian Arquiware signs deal to use new machine learning SDK to enhance sentiment analysis of social media networks.

La Jolla CA | São Paulo – Marketing consumer products in Brazil is about to get a lot easier thanks to a small, innovative software company. Arquiware is combining artificial intelligence with natural language processing to enable companies to analyze feelings and opinions social media networks. They are among the first in the world to apply techniques to create tools that enable companies to understand the sentiments of customers.

“Many multi-nationals come to Brazil then realize that it takes more than understanding Portuguese to understand how a brand interacts with consumers.” said Luis Lima, President of Arquiware, “In fact, the diversity of Brazil makes it almost impossible to understand our market unless you use sophisticated tools to extract the true meaning of what people are saying about you.”

Arquiware will add ai-one’s Topic-Mapper SDK to build artificial intelligence into two existing products. SentimentWare and TopicExtractWare analyze text data from social media networks and news feeds. The new capability gives Arquiware clients the ability to understand how any given news event will impact the perception of a brand.

ai-one’s technology is used by telecom companies, security and law enforcement agencies to enable computers to read text in a similar manner to humans. “We are thrilled that Arquiware will apply our technology to social media sentiment analysis. They are ahead of the efforts I have seen in Silicon Valley,” commented Olin Hyde, VP of Business Development for ai-one.

About Arquiware DSC (Brazil), Arquiware is one of Brazil’s leading software companies specializing in application development using natural language processing (NLP) and text mining. Arquiware builds custom applications for numerous enterprise clients and sells commercial off-the-shelf products for sentiment analysis and text extraction. SentimentWare provides sentiment analysis of social networks using the Radian6 API. TopicExtractWare is a SaaS that summarizes the meaning of any corpus of text by distilling information into tag clouds. Visit Arquiware’s free sentiment analysis of Twitter feeds to determine the best samba school in Carnival.

Contact:  Luis Lima Phone +55-11-233-82742, email: lglima (at) arquiware.com.br web: http://www.arquiware.com.br

###

 

Calling a support hotline?

July 13th, 2011

Then you might have IBM’s Watson helping you. So, maybe Star Trek’s talking computer is not so far off anymore, with Watson. IBM is looking to shop the technology around to industries like retail and customer service. Read more on Hemispheres Magazine.