è«æ
Twitterã§è¦ããã以ä¸ã®è¨äºã§ç´¹ä»ããã¦ããè«æãããããããã ã£ãã®ã§èªãã ææ³ã¨å 容ã®ã¦ãã¨ã¼ãªç´¹ä»(詳ããç¥ããã人ã¯å è«æãå¼ãã§ãã ãã)ãã¤ã¯ãã½ããã®ç 究è ã«ããããããã A/B ãã¹ãã«ããã¦ã¡ããªãã¯ã解éããã¨ãã«é¥ããã¡ãªâ¦
Machine Learning: The High Interest Credit Card of Technical Debt NIPS 2014ã®Workshopã®SE4ML: Software Engineering for Machine Learningã¨ããã®ã§çºè¡¨ãããè«æã£ã½ãã§ã Twitterã§è©±é¡ã«ãªã£ã¦ããã®ã§èªãã ã¡ã¢ã§ãä¹ ãã¶ãã«è±èªè«æãèªãã â¦
æ å ±å¦çå¦ä¼ 第214åèªç¶è¨èªå¦çç ç©¶ä¼ æ¦è¦ ãããä¸ã«ç»å ´ãããããªå´©ãã表è¨ã®ããã¹ãã®å½¢æ ç´ è§£æããããæ£è¦ã®è¡¨ç¾(å´©ãã¦ãªã表ç¾)ã¨å´©ãã表ç¾ã®ãã¢ã®ãã¼ã¿ãç¨æãã¦ãå´©ãæ¹ã®ãã¿ã¼ã³ã¨ä¿¡é ¼åº¦ãå¦ç¿ããå½¢æ ç´ è§£ææã«å©ç¨ãã ææ³ æ£è¦â¦
"Social Text Normalization using Contextual Graph Random Walks" (pdf) Twitterã¨ãã®ã½ã¼ã·ã£ã«ã¡ãã£ã¢ã§ã¯ãã ãã表ç¾ãå¤ãã®ã§ãããããããã¹ãã®æ£è¦åããã話 以ä¸è«æä¸ã®ä¾ã®ä¸é¨ wuz up bro (what is up brother) 4get (forget), 2morrow (â¦
Utopia/Microtopia Parallel Corpus âãã¼ã¿ãªã©ãå ¬éããã¦ããã¿ããã§ããã¤ã¯ãããã°(Twitterã¨Weibo)ä¸ã®ãã¤ã¼ãããããã©ã¬ã«ãª(翻訳ã«ãªã£ã¦ãã)ãæç« ãå«ã¾ãã¦ãããã®ãéãã¦ãã話 åºæ¬çã«ã¯ä¸ã¤ã®ãã¤ã¼ãä¸ã«è¤æ°ã®è¨èªãå«ã¾ãã¦ããâ¦
"Learning Latent Personas of Film Characters"(pdf) ãã£ã©ã¯ã¿ã¼ã®ã¿ã¤ã(personaãäººæ ¼)ãåæãã話ã èªç¶è¨èªå¦çã§ã¯ç©èªã®ããããã¨ãã¤ãã³ãã®é£éã®ç 究ã¯å¤ããã©ããã£ã©ã¯ã¿ã¼ã®ã¿ã¤ãã«å¯¾ããç 究ã¯åãã¦ãããæ ç»ã®ãã¼ã¿ã対象ã¨ãã¦ãâ¦
"Identification of Speakers in Novels"(pdf) å°èª¬ã®ã»ãªããã©ã®ç»å ´äººç©ã®ãã®ããæ¨å®ãã話ã æ師ããã®ã©ã³ãã³ã°å¦ç¿(SVM-rankãå©ç¨)ã 7å²ãããã¯æ£è§£ã§ãããããå è¡ç 究ã®æ師ããã®åé¡ã§ä½¿ããã¦ããç¹å¾´é(ã»ãªãã¨ã®è·é¢ãç»å ´äººç©ã®åºç¾â¦
"Exploiting Topic based Twitter Sentiment for Stock Prediction" æ¦è¦ Twitterããæ ªä¾¡ã®ææ¨ã®ä¸ä¸ãäºæ¸¬ãã話ã ä¼¼ããããªè©±ã¯åãããããã©ããã®è«æã§ã¯ãããã¯ã¢ãã«(åã®æ¥ã®ãããã¯ãèæ ®ããDirichlet Process Mixture)ã使ã£ã¦ããçµæãâ¦
Good, Great, Excellent: Global Inference of Semantic Intensities å³è¡¨ãå¼ã¯è«æä¸ããå¼ç¨ æ¦è¦ ä¼¼ãæå³ã®å½¢å®¹è©ã®å¼·å¼±ã®é åºä»ããããã¿ã¹ã¯ ä¾ãã°ãå¯ããã涼ããããå·ããããçãããªã©ã®å½¢å®¹è©ãä¸ããããæã«ããã®å¼·å¼±é¢ä¿ã¯ã涼ããï¼å¯â¦
"TopicSpam: a Topic-Model based approach for spam detection" æ¦è¦ spam detectionã¨ãªã£ã¦ããããã©ã以åç´¹ä»ãã以ä¸ã®è«æã¨åæ§ã«ãå½ã®ã¬ãã¥ã¼ã®æ¤åºãè¡ã£ã¦ãã è«æææ³: "Finding Deceptive Opinion Spam by Any Stretch of the Imagination"â¦
"The lie detector: explorations in the automatic recognition of deceptive language" ä¸ããããæç« ãåãã©ãããèå¥ããã¿ã¹ã¯Amazon Mechanical Turkã§å¤§å¢ã®äººã«ãæ¬å½ã®æè¦ã¨åã®æè¦ãæ¸ãã¦ããã£ã¦ãã¼ã¿ã»ããã«ãã¦ããã ãã¤ã¼ããã¤ãºãâ¦
è«æææ³: "Personalized PageRank vectors for tag recommendations: inside FolkRank" (RecSys 2011)
Personalized PageRank vectors for tag recommendations æ¦è¦ ã¦ã¼ã¶ã¼ã¨ã¢ã¤ãã ã¨ã¿ã°ã®ãã¼ã¿ãä¸ããããæã«ãã¦ã¼ã¶ã¼ã¨ã¢ã¤ãã ã«å¯¾ããã¿ã°ã®æ¨è¦ãè¡ãæ¹æ³ã¨ãã¦FolkRankã¨ããã¢ã«ã´ãªãºã ããã使ããã¦ãã(ããã)ã ãã®ã¢ã«ã´ãªãºã ãè¿ä¼¼â¦
SEXI 2013 | Workshop on Search and Exploration of X-Rated Information at WSDM 2013 WSDM (Web Search and Data Mining) 2013ã§SEXI(Search and Exploration of X-rated Information)ã¨ããã¯ã¼ã¯ã·ã§ãããéå¬ããã¦ä¸é¨ã§è©±é¡ã«ãªã£ã ãã®ã¯ã¼ã¯ã·ã§â¦
âRepresenting Topics Using Images", Nikolaos Aletras and Mark Stevenson ç 究室ã§è«æç´¹ä»ããã®ã§é©å½ã«ä»¥ä¸ã«è³æãè²¼ã£ã¦ããã è«æãèªãã å¾ã§æ°ã¥ãããã©ãNAACL 2013ã¯ã¾ã ãã£ã¦ããªãã®ã§ãpreprintãã¼ã¸ã§ã³(ï¼)ã£ã½ãã ä½æ ãç 究室ã§ã®â¦
"ããããããããããããããããããï¼ï¼ï¼ï¼ï¼ï¼ï¼ï¼ãã¤ã¯ãããã°ãç¨ããæ師ãªãå«åãã¬ã¼ãºæ½åº"(DEIM 2013) ã¿ã¤ãã«ãæ°ã«ãªã£ãã®ã§èªã¿ã¾ãããäºç¨¿ã§ãã DEIM2013 æ«å®äºç¨¿é æ¦è¦ å 容ã¯ããããããããããã¨ãã¿ã¤ãã«ã«ãããããããâ¦
è«æææ³: "TCSST: Transfer Classification of Short & Sparse Text Using External Data"(CIKM 2012)
TCSST: Transfer Classification of Short & Sparse Text Using External Data æ¦è¦ Twitterãªã©ã®ãã¤ã¯ãããã°ãæ¤ç´¢ã¨ã³ã¸ã³ã®ã¹ãããããã¬ãã¥ã¼ãªã©ã®çãã¹ãã¼ã¹ãªæç« ã«å¯¾ãã¦ãå¤é¨ã®ãªã½ã¼ã¹ã使ã£ã¦è»¢ç§»å¦ç¿ãè¡ããã¨ã§ç²¾åº¦ã®åä¸ããããé¢â¦
Creating Stories: Social Curation of Twitter Messages Akisato Kimura: Research Interests: Assisting social curation of Twitter messages è¿½è¨ èªãã è«æã¨ãªã³ã¯å ã®è«æ(4ãã¼ã¸)ãå¾®å¦ã«éãã£ã½ãã¨ããããåã«å°å·ãããã®ã¨åãå 容ã®è«æ(8â¦
æ¦è¦ 1996å¹´ã®ããªãå¤ãè«æã é ããã«ã³ãã¢ãã«(Hidden Markov Model, HMM)ã§ä¸å½èªã®å½¢æ ç´ è§£æ(åèªåå²ã¨åè©ã¿ã°ä»ã)ããã話ã ææ³ ä¸å½èªãæ¥æ¬èªãªã©ã®åãã¡æ¸ãããã¦ããªãè¨èªã§ãHMMã使ã£ã¦åèªåå²ã¨åè©ã¿ã°ä»ããããå ´åã®å¼ãç¥ããâ¦
"Automatically Constructing a Normalisation Dictionary for Microblogs" æ¦è¦ Twitterã¨ãã§ã®åèªã®æ£è¦åç¨ã®è¾æ¸ãä½ã話ã ä¾ã2morwâtomorrow ææ³ æèé¡ä¼¼åº¦ã®ä¼¼ããè¾æ¸ã«ãªãåèª(OOV)ã¨è¾æ¸ã«ããåèª(IVã®ãã¢)ãéãã éãããã¢ãæååçâ¦
"Adversarial Support Vector Machine Learning" æ¦è¦ æµå¯¾çå¦ç¿ã®è©±ã spamæ¤åºãä¾µå ¥æ¤åºãªã©ã®ç¸æããã§ãã¯ãããããããã¨ãã¦é常ã«è¦ãããããã¨ããå ´åã®æ©æ¢°å¦ç¿ã See: æµå¯¾çå¦ç¿ - æ©æ¢°å¦ç¿ã®ãæ±é·ºã®æWikiã å è¡ç 究ã§ã¯æªæã®ããæµå¯¾â¦
Finding Bursty Topics from Microblogs æ¦è¦ ãããã¯æ¨å®ã§ç¨ãããããã¨ã®å¤ãLDA(Latent Dirichlet Allocation)ã«ä»¥ä¸ã®2ã¤ã®ãã¤ã¯ãããã°ç¹æã®è¦ç´ ãå ããï¼ åãæéã«ã¯åããããªãããã¯ãåºãããï¼ä¾ãã°å¤§äºä»¶ãèµ·ããæ¥ ã¦ã¼ã¶ã¼ã¯æéã«â¦
Deciphering Foreign Language by Combining Language Model and Context Vectors æ¦è¦ é常ã®çµ±è¨çæ©æ¢°ç¿»è¨³ã§ã¯ãã©ã¬ã«ã³ã¼ãã¹ã¨ããåãæã®å¯¾è¨³ãã¼ã¿ãã³ã³ãã©ãã«ã³ã¼ãã¹ã¨ããåããããªãããã¯ã«ã¤ãã¦æ¸ããããã¼ã¿ãªã©ãå©ç¨ãããã¨ãå¤ãâ¦
Stylometric Analysis of Scientific Articles æ¦è¦ è«æã®æä½ãã以ä¸ã®3ã¤ã®ã¿ã¹ã¯ãSVMãç¨ãã¦èå¥ï¼ è±èªãã¤ãã£ããå¦ã æ§å¥ ã«ã³ãã¡ã¬ã³ã¹ãã¯ã¼ã¯ã·ã§ããã 使ç¨ããã¦ããfeature Bow(åèª) Style(æä½) Syntax(æ§æ) çµæ Få¤ã§ãã¤ãã£ããâ¦
Word Salad: Relating Food Prices and Descriptions æ¦è¦ ã¬ãã¥ã¼æãã極æ§(ãã¸ãã£ãããã¬ãã£ãã)ãæ¨å®ããã£ã¦è©±ã¯ããããããããã©ï¼ããã¯èª¬ææãã¬ãã¥ã¼æããé£ã¹ç©ã®å¤æ®µãæ¨å®ããã£ã¦ããé¢ç½ãã¿ã¹ã¯ï¼ç´ æ§ã¨ãã¦ã¯ã¡ãã¥ã¼åã説ææâ¦
Wiki-ly Supervised Part-of-Speech Tagging æ¦è¦ ããªã¼ãªè¾æ¸ã®Wiktionaryãå©ç¨ãã¦ãå¼±æ師ããå½¢æ ç´ è§£æãè¡ããWiktionaryã«ç»é²ããã¦ããåè©ãé¸ã°ããããã«å¶éãããã¦ï¼é ããã«ã³ãã¢ãã«ãEMã¢ã«ã´ãªãºã ãç¨ãã¦å¦ç¿ï¼Unsupervisedãªå ´åãâ¦
An Empirical Investigation of Statistical Significance in NLP æ¦è¦ èªç¶è¨èªå¦çã«ãããæææ§æ¤å®ã«é¢ãã調æ»ï¼ ææ³ ããã¤ãã®ã¯ã¼ã¯ã·ã§ãããshared taskã§ã¯åãã¼ã ã®åºåãå ¬éããã¦ãã¦ï¼ãããã«å¯¾ãã¦æ¤å®ãè¡ãªã£ã¦è©ä¾¡ææ¨ã¨p-valueã®é¢â¦
Multi-Domain Learning: When Do Domains Matter? æ¦è¦ ãã«ããã¡ã¤ã³ã©ã¼ãã³ã°(MDL)ã«ãããããã¤ãã®çåã«å¯¾ããæ¤è¨¼ï¼ MDLã«ãã精度ä¸æã¯å¿ ããããã¡ã¤ã³ã®å½±é¿ã§ã¯ãªãï¼ã¢ã³ãµã³ãã«ã®å¹æã§ã¯ãªããï¼ ãã¡ã¤ã³ã©ãã«ãã©ã³ãã ã«å ¥ãæ¿ãã¦ãâ¦
Polarity Inducing Latent Semantic Analysis æ¦è¦ ææ¸-åèªè¡åä¸ã§cosé¡ä¼¼åº¦ãªã©ã使ãã¨ï¼æå³çã«ä¼¼ãåèªå士ãè¦ã¤ãããã¨ãã§ããï¼ ããããã®å ´åã®ãæå³çã«ä¼¼ããã¨ããã®ã¯ãé¡ç¾©èªãããã¯åæèªãã§ããï¼ãåæèªãã ããè¦ã¤ãããå ´åã«â¦
"Exploring Topic Coherence over many models and many topics" æ¦è¦ ãããã¯ã¢ãã«[LSA(SVDã«ãã), LSA(NMFã«ãã), LDA]ã®æ¯è¼ãè¡ãã åºåãããããã¯ã®ã³ãã¼ã¬ã³ã¹(ä¸è²«æ§)ãè¿å¹´ææ¡ãããææ³(UCI measure, UMass measure)ã®å¹³åãã¨ã³ãããã¼â¦
http://www.cs.washington.edu/homes/aritter/mt_chat.pdf ç 究室ã®è«æç´¹ä»ã§ç´¹ä»ãããEMNLP2011ã®è«æã æ¦è¦ Twitterãªã©ã®ãããªã¡ãã»ã¼ã¸ã«å¯¾ããå¿çãèªåçæã ã¡ãã»ã¼ã¸ã¨ãã®å¿çã®ã³ã¼ãã¹ã«å¯¾ãã¦çµ±è¨çæ©æ¢°ç¿»è¨³ã®ææ³ã使ã£ãããé¡ä¼¼åº¦ã®â¦