Comment le web de données change-t-il la nature de la toile ? En rendant les contenus du web lisibles par les machines, le web sémantique bouleverse notre univers informationnel et ouvre de nouvelles opportunités propres à redéfinir la nature du Web : d’un web de document à un web de données. (ce billet est issue d’une note de synthèse, réalisée dans le cadre de mes activités universitaires. Il s’agit d’un bilan de lecture autour du web de données. Il m’a semblé intéressant de le republier ici pour solliciter l’avis des connaisseurs de ce sujet, et ouvrir le débat) 1. Croissance exponentielle du volume et de la valeur des données : le terreau d’éclosion du web de données A peine avons-nous commencé à explorer les nouveaux modèles d’affaires du Web 2.0 que déjà se profile un nouveau paradigme prometteur : le web de données. Les applications du Web 2.0 reposent de plus en plus sur la gestion, l’analyse et l’exploitation des massives quantités de données issues des UGC. 2. information overload ("seven months" by dylanroscover) 3. 4. 5. 6.
Named-entity recognition Named-entity recognition (NER) (also known as entity identification, entity chunking and entity extraction) is a subtask of information extraction that seeks to locate and classify elements in text into pre-defined categories such as the names of persons, organizations, locations, expressions of times, quantities, monetary values, percentages, etc. Most research on NER systems has been structured as taking an unannotated block of text, such as this one: Jim bought 300 shares of Acme Corp. in 2006. And producing an annotated block of text that highlights the names of entities: [Jim]Person bought 300 shares of [Acme Corp.]Organization in [2006]Time. In this example, a person name consisting of one token, a two-token company name and a temporal expression have been detected and classified. State-of-the-art NER systems for English produce near-human performance. Problem definition[edit] Certain hierarchies of named entity types have been proposed in the literature. Formal evaluation[edit]
Stanford shows off their federated search tool » Federated Searc 26Jan Blog sponsor Deep Web Technologies built a federated search tool for Stanford University. I was involved with the first prototype and I’m proud of what the Stanford/Deep Web Technologies partnership has accomplished. My involvement with the Stanford federated search tool was multi-faceted. Tags: federated search
World Service Radio Archive Prototype Semanlink Home Page Mafait.org - project Thinknowlogy - Fundamentally designed Artificial Intelligence How-To: Search the Social Web – Ultimate Toolkit Are you using content marketing as part of your digital strategy to grow your business? If so, you're not alone. According to the Content Marketing Institute, the lion's share of marketers (some 92%) report using content marketing. In the fast moving world of digital strategy, things are always changing. What should you expect in 2014 to change in the world of content marketing? Hana Abaza of Uberflip has put together an infographic detailing five key content marekting trends for the coming year. 1. 2. 3. 4. 5.
Free Computer, Programming, Mathematics, Technical Books, Lecture Notes and Tutorials Synaptic Web Stay updated about the Synaptic Web on Twitter via @SynapticWeb The Synaptic Web By Khris Loux, Eric Blantz, Chris Saad and you... The Internet is constantly evolving. As the speed, flexibility and complexity of connections increase exponentially, the Web is increasingly beginning to resemble a biological analog; the human brain. But what exactly is it that’s makes us, or the Web, smart? In the brain, neurologists now believe that it is the density and flexibility of the connections between neurons, not simply neurons themselves, which are at the root of intelligence. Even if the total number of brain cells, or neurons, begins to diminish in early adulthood, our ability to generate new connections between neurons and between different parts of the brain – what neurologist call “plasticity” - persists throughout life. Signs of the emerging Synaptic Web abound. The same is true for Social Networks. Social profiles are becoming real-time streams. Databases are becoming data peers.
LingPipe Home How Can We Help You? Get the latest version: Free and Paid Licenses/DownloadsLearn how to use LingPipe: Tutorials Get expert help using LingPipe: Services Join us on Facebook What is LingPipe? LingPipe is tool kit for processing text using computational linguistics. LingPipe is used to do tasks like: Find the names of people, organizations or locations in newsAutomatically classify Twitter search results into categoriesSuggest correct spellings of queries To get a better idea of the range of possible LingPipe uses, visit our tutorials and sandbox. Architecture LingPipe's architecture is designed to be efficient, scalable, reusable, and robust. Latest Release: LingPipe 4.1.2 Intermediate Release The latest release of LingPipe is LingPipe 4.1.2, which patches some bugs and documentation. Migration from LingPipe 3 to LingPipe 4 LingPipe 4.1.2 is not backward compatible with LingPipe 3.9.3. Programs that compile in LingPipe 3.9.3 without deprecation warnings should compile and run in Lingpipe 4.1.2.
Bing Goes The iPhone. Still Great For Porn. Since the dawn of Bing, it’s been exceptionally good at one thing: Finding porn. Its new iPhone app, which launched tonight in the App Store, is no different. By default, the app has a Safe Search setting of “Moderate.” Searching for “porn” this way yields several promising results. However, with just two clicks, any kids can turn off safe search and off they go! I love this for two reasons: 1) The app is rated 4+, yet it’s super simple to gain access to hardcore porn in a few clicks. To be fair, Google’s iPhone app also allows you to search for porn. All that said, the Bing app is actually quite nice. Both images below taken on a search for “porn” with safe search turned off. 700 Free Movies Online: Great Classics, Indies, Noir, Westerns Watch 4,000+ movies free online. Includes classics, indies, film noir, documentaries and other films, created by some of our greatest actors, actresses and directors. The collection is divided into the following categories: Comedy & Drama; Film Noir, Horror & Hitchcock; Westerns (many with John Wayne); Martial Arts Movies; Silent Films; Documentaries, and Animation. Free Comedy & Dramas 125 Korean Feature Films — Free — The Korean Film Archive has put on YouTube over 100 Korean feature films, including Im Kwon-taek’s Sopyonje and Hong Sangsoo’s The Day the Pig Fell Into a Well. Free Hitchcock, Noir, Horror & Thriller Films A Bucket of Blood - Free — Roger Corman’s classic comedy/horror film set in Bohemian San Francisco. Find a complete collection of Film Noir movies here and Alfred Hitchcock movies here. Free Kung Fu & Martial Arts Films
Web 3.0 A short story about the Semantic Web. Some Internet experts believe the next generation of the Web - Web 3.0 - will make tasks like your search for movies and food faster and easier. Instead of multiple searches, you might type a complex sentence or two in your Web 3.0 browser, and the Web will do the rest. For example, you could type "I want to see a funny movie and then eat at a good Mexican restaurant. What are my options?" That's not all. Eventually you might be able to ask your browser open questions like "where should I go for lunch?" Watch the full documentary now
AutoMap: Project Overview | People | Sponsors | Publications | Hardware Requirements | Software | Training & Sample Data AutoMap is a text mining tool developed by CASOS at Carnegie Mellon. Input: one or more unstructured texts. Output: DyNetML files and CS files. AutoMap enables the extraction of information from texts using Network Text Analysis methods. AutoMap exists as part of a text mining suite that includes a series of pre-processors for cleaning the raw texts so that they can be processed and a set of post-processor that employ semantic inferencing to improve the coding and deduce missing information. AutoMap uses parts of speech tagging and proximity analysis to do computer-assisted Network Text Analysis (NTA). AutoMap subsumes classical Content Analysis by analyzing the existence, frequencies, and covariance of terms and themes. AutoMap has been implemented in Java 1.7. It can operate in both a front end with gui, and backend mode. Main functionalities of AutoMap are: "From Texts to Networks"