]> 2014-04-24T01:05:58Z Maximum Entropy Modeling 2014-04-24 Collection of links, papers, software... 2014-04-05T17:43:46Z L'extinction de masse du Permien est-elle due à des micro-organismes ? 2014-04-05 EC-Web'14: The 15th International Conference on Electronic Commerce and Web Technologies 2014-04-18T23:35:51Z 2014-04-18 2014-04-23T22:03:44Z 2014-04-23 nltk - hierarchical classification + topic model training data for internet articles and social media - Stack Overflow 2014-04-07T16:58:26Z 2014-04-07 A good man in Rwanda "Why we fear Google", Mathias Döpfner’s open letter to Eric Schmidt 2014-04-26 2014-04-26T11:41:33Z Topic Modeling and Network Analysis | the scottbot irregular 2014-04-23T22:51:15Z 2014-04-23 Great post 2014-04-23T10:56:23Z topic-modeling-tool - A graphical user interface tool for topic modeling - Google Project Hosting 2014-04-23 2014-04-22 pallet - A professionalization of the UMass project "Mallet" - Google Project Hosting 2014-04-22T17:39:23Z An Introduction to Conditional Random Fields for Relational Learning (Charles Sutton and Andrew McCallum, 2006) 2014-04-24 2014-04-24T01:16:48Z Daniel Cohn-Bendit insiste, une dernière fois, sur la «nécessité d’Europe » 2014-04-17T13:55:35Z 2014-04-17 2014-04-08T19:18:28Z 2014-04-08 Machine Learning Tutorial: The Max Entropy Text Classifier | DatumBox java - Multi-Label Document Classification - Stack Overflow 2014-04-25T19:22:48Z 2014-04-25 « Quelques-uns l’ont voulu, d’autres l’ont fait, tous l’ont laissé faire. » ("A shocking crime was committed on the unscrupulous initiative of few individuals, with the blessing of more, and amid the passive acquiescence of all") (Tacite)<br/> « S’il y a eu un génocide, il peut y en avoir un autre, puisque la cause est toujours là et qu’on ne la connaît pas. » (rapporté par Jean Hatzfeld) 2014-04-05 2014-04-05T18:37:31Z Comment devient-on un bourreau ? Why No One Trusts Facebook To Power The Future – ReadWrite 2014-04-05T23:36:53Z 2014-04-05 people need to know that Facebook is making things to improve the human experience, not just spending billions to make even more billions off our personal information By the authors of LIBSVM 2014-04-25 2014-04-25T13:33:06Z A practical guide to Support Vector classification 2014-04-13 2014-04-13T10:21:23Z "Na am Francophonie" Sogha Niger - YouTube Caitalism 3.0 - A guide to reclaiming the commons 2014-04-23T21:47:22Z 2014-04-23 2014-04-16 2014-04-16T01:08:36Z Nick Grossman's Slow Hunch — Should we regulate the Internet the real world way or the real world the Internet way? Comment les sites de commerce nous manipulent 2014-04-13T19:33:43Z 2014-04-13 All About Bitmap Indexes... And Sorting Them 2014-04-23T21:48:30Z 2014-04-23 2014-04-21 2014-04-21T12:20:28Z GoodRelations: technical report 2014-04-23 2014-04-23T21:52:03Z Dans les vergers du Sichuan, les hommes font le travail des abeilles 2014-04-02T18:36:43Z 2014-04-02 Digit recognition to solve Sudoku puzzles automatically with a webcam - MATLAB Vidéo 2014-04-19 2014-04-19T10:49:21Z OPDM - Ontologies 2014-04-30 Data Virtuality 2014-04-30T17:17:43Z 2014-04-30T13:59:32Z 2014-04-30 Stanford bioengineers create circuit board modeled on the human brain | Stanford News Release 10 Tips to Improve your Text Classification Algorithm Accuracy and Performance | Thinknook 2014-04-07T10:13:59Z 2014-04-07 Learn to Say “I Dont Know" 2014-04-24 2014-04-24T00:00:04Z machine learning - Unsupervised automatic tagging algorithms? - Stack Overflow Re: Generic Property-Value Proposal for Schema.org from Francois-Paul Servant on 2014-04-30 (public-vocabs@w3.org from April 2014) 2014-04-30T17:20:29Z 2014-04-30 2014-04-23 2014-04-23T21:58:12Z Thomas Piketty : « Le retour des inégalités inquiète aux Etats-Unis » 2014-04-23 In this paper, we formally define the problem of topic modeling with network structure (TMN). We propose a novel solution to this problem, which regularizes a statistical topic model with a harmonic regularizer based on a graph structure in the data. The proposed method combines topic modeling and social network analysis, and leverages the power of both statistical topic models and discrete regularization. The output of this model can summarize well topics in text, map a topic onto the network, and discover topical communities. Topic modeling with network regularization 2014-04-23T10:54:41Z Les fossoyeurs de l’innovation | L'Âge de la multitude 2014-04-11T23:47:28Z 2014-04-11 Text classification using string kernels 2014-04-08T18:46:34Z 2014-04-08 2014-04-08 2014-04-08T18:50:37Z Text categorization - Scholarpedia 2014-04-08 Inductive learning algorithms and representations for text categorization 2014-04-08T19:08:52Z 2014-04-22 2014-04-22T18:21:08Z Real-Time Topic Modeling of Microblogs 2014-04-08 2014-04-08T16:24:34Z [java-nlp-user] Stanford NER: confidence scores Sapping Attention: When you have a MALLET, everything looks like a nail 2014-04-25T12:44:00Z 2014-04-25 How can Neural Networks be applied to Time Series Forecasting? 2014-04-28T15:40:13Z 2014-04-28 2014-04-13 2014-04-13T13:40:35Z Microdata and RDFa Living Together in Harmony | Jeni's Musings 2014-04-30T13:23:01Z "in the absence of a good description of schema.org I tried to put one together for my own purposes, as a potential producer and consumer of schema.org information. This ended up needing to include a philosophy of just what schema.org is, so I put in my own, namely something that could serve as a precursor to a formal treatment of schema.org" schema.org as it could be from Peter F. Patel-Schneider on 2014-01-06 (public-vocabs@w3.org from January 2014) 2014-04-30 Machine Learning for Sequential Data: A Review 2014-04-28 2014-04-28T15:56:10Z 2014-04-08 2014-04-08T16:53:07Z java - Method(s) to output confidence score from Stanford Classifier? - Stack Overflow from the abstract:<br/> Modern machine learning algorithms can extract useful information from text, images and videos. All these applications involve solving NP-hard problems in average case using heuristics. What properties of the input allow it to be solved effciently? Theoretically analyzing the heuristics is very challenging. Few results were known. <br/> This thesis takes a different approach: we identify natural properties of the input, then design new algorithms that provably works assuming the input has these properties. We are able to give new, provable and sometimes practical algorithms for learning tasks related to text corpus, images and social networks. <br/> ...In theory, the assumptions in this thesis help us understand why intractable problems in machine learning can often be solved; in practice, the results suggest inherently new approaches for machine learning. 2014-04-23 2014-04-23T22:21:47Z Provable Algorithms for Machine Learning Problems by Rong Ge. Learning Multilabel classification of news articles (2013) 2014-04-08T17:20:45Z 2014-04-08 > The notion of ’tip-off’ words (words that are highly indicative of the article belonging to a particular topic) suggested to us that fairly robust multi-label classification should be achievable with only a limited set of high-information words, and moreover, without access to any explicit priors on class labels > On the whole our research validated the common approach of using binary-classifiers to learn multi-label topic classifications for new articles. The tfidf approach captures some interesting aspects of the intuition behind how people may classify news articles, but we were not able to lower the error produced by the tfidf model sufficiently to make it practically competitive with the binary classification scheme Bing - Knowledge Widget (Beta) 2014-04-04T13:19:32Z 2014-04-04 On Bayesian inference, maximum entropy and Support Vector Machines methods 2014-04-25T16:21:40Z 2014-04-25 Robobees 2014-04-29T01:29:55Z 2014-04-29 2014-04-30 IEEE Xplore Abstract - Neurogrid: A Mixed-Analog-Digital Multichip System for Large-Scale Neural Simulations 2014-04-30T14:01:36Z n this paper, we describe the design of Neurogrid, a neuromorphic system for simulating large-scale neural models in real time. Neuromorphic systems realize the function of biological neural systems by emulating their structure. Designers of such systems face three major design choices: 1) whether to emulate the four neural elements—axonal arbor, synapse, dendritic tree, and soma—with dedicated or shared electronic circuits; 2) whether to implement these electronic circuits in an analog or digital manner; and 3) whether to interconnect arrays of these silicon neurons with a mesh or a tree network. The choices we made were: 1) we emulated all neural elements except the soma with shared electronic circuits; this choice maximized the number of synaptic connections; 2) we realized all electronic circuits except those for axonal arbors in an analog manner; this choice maximized energy efficiency; and 3) we interconnected neural arrays in a tree network; this choice maximized throughput. These three choices made it possible to simulate a million neurons with billions of synaptic connections in real time—for the first time—using 16 Neurocores integrated on a board that consumes three watts. How to do text classification with label probabilities? - Stack Overflow 2014-04-25 2014-04-25T19:10:57Z P vs. NP Problem Linked To the Quantum Nature of the Universe - Slashdot 2014-04-05T23:55:46Z 2014-04-05 2014-04-18 2014-04-18T12:02:39Z Le Brésil va lâcher des millions de moustiques OGM contre la dengue | Eco(lo) 2014-04-07 Augusta Cantwell’s boss wants a new slideshow. Will Daria Norton be of any help? 2014-04-07T00:00:05Z Data is the new NEW — WHAT? Efficient Multi-label Classification with Many Labels (2013) 2014-04-25T19:21:16Z 2014-04-25