Semantic Web Takes Off
SUMMARYWhile the early days of the Web were all about having access to the billions of documents that could be discovered using search engines via keyword analysis, Web 2.0/social media is all about identifying connections between people - these connections are all a part of the a lager network between e-mails, blogs, social networks etc. sometimes referred to as the ‘social graph’. By harnessing the "wisdom of crowds" by adding metadata to Web pages (using tags, ratings etc), it has become easier to search, locate and organize information in a systematic manner (sites like Digg, Flickr, Del.icio.us, eBay). Since it would be almost impossible to re-structure all the data by manually insert metadata in the billions of existing Web pages, the semantic Web is making it easier for us to locate information by enabling computers to do the thinking and interpret the meanings of Web pages and determine the connections between them. Though we may be far away from machines doing the all the thinking/work for us, a lot of research is ongoing and several startups like Twine, Freebase, Hakia, Spock, Clear Forest and TripIt are testing the use of semantic Web technologies to develop smarter ways to collaborate, find, share and combine information.
THE PLAY
A simple explanation of the semantic Web (machine learning/AI) is the idea of representing information in a structured form so that computers can understand it and then solve complex problems. Today there are a number of sites that have several elements of the semantic Web. Wikipedia presents relevant information in a structured format divided into sub-topics/categories with relevant external links/references. MySpace and Facebook represent connections/links among people being part of a network. Amazon and Blockbuster recommendation systems presents users with a list of alternatives of books, movies or music based on the user’s past interest/behavior (product navigation/purchase behavior). These examples are pretty simple comparred to what the Semantic Web is diving into. With the semantic Web a whole new experience is about to be delivered. For instance, imagine asking the computer to come up with a list of travel alternatives for a week’s vacation with a budget under $5000 for a couple or asking to list down prices of smartphones with screen size larger than 3 inches with 8GB memory at the nearest shops that is open until 8pm -- answering these types of complex questions is the underlying challenge of the future Web.
There’s a growing interest among researchers, scientists and tech startups to bring a new level of intelligence to the Web. A startup called Twine allows users to place information that’s important to them, be it emails, Web searches, videos, calendars, notes, documents, spreadsheets etc, onto the site. Twine then starts to analyze the information and automatically sort it into categories that include people involved, concepts discussed, projects, places etc, making it easier to search, organize, collaborate and share in a smart way. It also has elements of social networking where users can have access to information collected by others in their network. The site uses advanced machine learning and natural-language techniques as well as the combination of social connections to automatically tag data. Later this year, it plans to open its platform to software developers to build tools such as visualization software so that users can see information in different ways.
TripIt makes it easier to manage your travel plans. Travelers are required to forward all their booking confirmation emails (flights, hotels, car rentals etc.) to plans@tripit.com and the “itinerator” technology would then extract useful information from these mails, presenting it in a single structured and organized itinerary. The system then adds information from the Web for places that you visit which includes weather, local maps, pictures etc. Travelers can access their itineraries on their mobile device or sync the details on the personal calendar to share it with fellow travelers. Hakia and Powerset attempt to analyze sentences/questions based on the understanding of natural language and return possible answers to those queries.
Decades of research/experiments in artificial intelligence are just beginning to take off. It remains to be seen how far humans can go in transferring intelligence that will enable machines to do the thinking for us.
Resources: Technology Review, Read/Write Web
Bob Rattivarakorn
Trends Research Analyst
RarePlay
