ICDM SENTIRE


Sentiment Elicitation from Natural Text for Information Retrieval and Extraction (SENTIRE) is the IEEE ICDM workshop series on opinion mining. The term SENTIRE comes from the Latin feel and it is root of words such as sentiment and sensation. SENTIRE aims to provide an international forum for researchers in the field of opinion mining and sentiment analysis to share information on their latest investigations in social information retrieval and their applications both in academic research areas and industrial sectors. The broader context of the workshop comprehends Web mining, AI, Semantic Web, information retrieval and natural language processing. For more info, write to workshop_name@sentic.net

SENTIRE'16 (ICDM 2016, December 12th, Barcelona)
SENTIRE'15 (ICDM 2015, November 14th, Atlantic City)
SENTIRE'14 (ICDM 2014, December 14th, Shenzhen)
SENTIRE'13 (ICDM 2013, December 7th, Dallas)
SENTIRE'12 (ICDM 2012, December 10th, Brussels)
SENTIRE'11 (ICDM 2011, December 11th, Vancouver)

RATIONALE
Memory and data capacities double approximately every two years and, apparently, the Web is following the same rule. User-generated contents, in particular, are an ever-growing source of opinion and sentiments which are continuously spread worldwide through blogs, wikis, fora, chats and social networks. The distillation of knowledge from such sources is a key factor for applications in fields such as commerce, tourism, education and health, but the quantity and the nature of the contents they generate make it a very difficult task. Due to such challenging research problems and wide variety of practical applications, opinion mining and sentiment analysis have become very active research areas in the last decade.

Our understanding and knowledge of the problem and its solution are still limited as natural language understanding techniques are still pretty weak. Most of current research in sentiment analysis, in fact, merely relies on machine learning algorithms. Such algorithms, despite most of them being very effective, produce no human understandable results such that we know little about how and why output values are obtained. All such approaches, moreover, rely on syntactical structure of text, which is far from the way the human mind processes natural language. Next-generation opinion mining systems should employ techniques capable to better grasp the conceptual rules that govern sentiment and the clues that can convey these concepts from realization to verbalization in the human mind.

TOPICS
SENTIRE aims to provide an international forum for researchers in the field of opinion mining and sentiment analysis to share information on their latest investigations in social information retrieval and their applications both in academic research areas and industrial sectors. The broader context of the workshop comprehends Web mining, AI, Semantic Web, information retrieval and natural language processing. Topics of interest include but are not limited to:
• Sentiment identification & classification
• Opinion and sentiment summarization & visualization
• Aspect extraction for opinion mining
• Linguistic patterns for sentiment analysis
• Learning word dependencies in text
• Statistical learning theory for big social data analysis
• Deep learning for sarcasm detection
• Sentic computing
• Large commonsense graphs
• Conceptual primitives for sentiment analysis
• Multimodal emotion recognition and sentiment analysis
• Time evolving opinion & sentiment analysis
• Semantic multidimensional scaling for sentiment analysis
• Multidomain & cross-domain evaluation
• Domain adaptation for sentiment classification
• Affective knowledge acquisition for sentiment analysis
• Sentiment topic detection & trend discovery
• Social network analysis
• Social media marketing
• Opinion spam detection


SENTIRE'16 (ICDM 2016, December 12th, Barcelona)


PROGRAM
09:30-09:40 Welcoming and introduction (Erik Cambria)
09:40-10:30 Keynote - Harnessing reviews to build richer models of opinions (Julian McAuley)

10:30-11:00 Coffee break

11:00-11:25 Scalable and Real-time Sentiment Analysis of Twitter Data (Maria Karanasou)
11:30-11:55 Mining the Opinionated Web: Classification and Detection of Aspect Contexts for Aspect Based Sentiment Analysis (Oscar Araque)
12:00-12:25 Seasonal Fluctuations in Collective Mood Revealed by Wikipedia Searches and Twitter Posts (Nello Cristianini)
12:30-12:55 SmartVideoRanking: Video Search by Mining Emotions from Time-Synchronized Comments (Kosetsu Tsukuda)

13:00-14:30 Lunch break

14:30-14:50 Research On Sentiment Analysis: The First Decade (Oskar Ahlgren)
14:55-15:15 Bayesian Deep Convolution Belief Networks for Subjectivity Detection (Iti Chaturvedi)
15:20-15:40 Multi-Sentiment Modeling with Scalable Systematic Labeled Data Generation via Word2Vec Clustering (Dhruv Mayank)
15:45-16:05 Sentiment Lexica from Paired Comparisons (Christoph Dalitz)

16:10-16:30 Coffee break

16:30-16:50 Lexicon Knowledge Extraction with Sentiment Polarity Computation (Zhaoxia Wang)
16:55-17:15 The Truth and Nothing but the Truth: Multimodal Analysis for Deception Detection (Mimansa Jaiswal)
17:20-17:40 Modeling Satire in English Text for Automatic Detection (Aishwarya Reganti)
17:45-18:05 What Drives Consumer Choices? Mining Aspects and Opinions on Large Scale Review Data using Distributed Representation of Words (Linda Petzold)
18:05-18:10 Concluding remarks (Erik Cambria)

SPEAKER
Dr. McAuley has been an Assistant Professor in the Computer Science Department at the University of California, San Diego since 2014. Previously, he was a postdoctoral scholar at Stanford University after receiving his PhD from the Australian National University in 2011. His research is concerned with developing predictive models of human behavior using large volumes of online activity data.

KEYNOTE
Online reviews are often our first port of call when considering products and purchases online. Yet navigating huge volumes of reviews (many of which we might disagree with) is laborious, especially when we are interested in some niche aspect of a product. This suggests a need to build models that are capable of capturing the complex and idiosyncratic semantics of reviews, in order to build richer and more personalized recommender systems. In this talk I'll discuss three such directions: First, how can reviews be harnessed to better understand the dimensions (or facets) of people's opinions? Second, how can reviews be used to answer targeted questions, that may be subjective or require personalized responses? And third, how can reviews themselves be synthesized, so as to predict what a reviewer would say, even for products they haven't seen yet?

ORGANIZERS
• Erik Cambria, Nanyang Technological University (Singapore)
• Bing Liu, University of Illinois at Chicago (USA)
• Amir Hussain, University of Stirling (UK)
• Yongzheng Zhang, LinkedIn Inc. (USA)

sentire2016


SENTIRE'15 (ICDM 2015, November 14th, Atlantic City)


PROGRAM
14:00-14:05 Welcoming and introduction
14:05-14:45 Invited Talk (Chang-Tien Lu)
14:45-15:05 An Ensemble Sentiment Classification System of Twitter Data for Airline Services Analysis (Yun Wan)
15:05-15:25 Production Estimation for Shale Wells with Sentiment-based Features from Geology Reports (Bin Tong)
15:25-15:45 OntoSeg: a Novel Approach to Text Segmentation using Ontological Similarity (Mostafa Bayomi)

15:45-16:00 Coffee break

16:00-16:20 Improving Out-of-domain Sentiment Polarity classification using Argumentation (Lucas Carstens)
16:20-16:40 Exploiting the Focus of the Document for Enhanced Entities’ Sentiment Relevance Detection (Zvi Ben-Ami)
16:40-17:00 Towards Domain-Independent Opinion Target Extraction (Aleksander Wawer)

17:00-17:20 Lexical Resource for Medical Events: A Polarity Based Approach (Anupam Mondal)
17:20-17:40 Sentiment Polarity Classification using Structural Features (Daniel Ansari)
17:40-17:45 Concluding remarks

SPEAKER
Chang-Tien Lu is an Associate Professor of the Department of Computer Science and Associate Director of the Discovery Analytics Center at Virginia Tech. He served as Program Co-Chair of the 18th IEEE International Conference on Tools with Artificial Intelligence in 2006 and as General Co-Chair of the 20th IEEE International Conference on Tools with Artificial Intelligence in 2008 and the 17th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems in 2009. He also served as Secretary (2008-2011) and Vice Chair (2011-2014) of the ACM Special Interest Group on Spatial Information (ACM SIGSPATIAL). Dr. Lu's research work focuses on emerging requirements for analyzing, retrieving, and visualizing massive data. His ongoing projects range from explorations of fundamental access issues to practical applications that deal with data analysis and knowledge discovery tasks. His research has been sponsored by the NSF, NIH, DoD, IARPA, VDOT, and DCDOT. He received his Ph.D. in Computer Science from the University of Minnesota at Twin Cities.

KEYNOTE
Social media has become a popular data source as a surrogate for monitoring and detecting events. Analyzing social media (e.g., tweets) to reveal event information requires sophisticated techniques. Tweets are written in unstructured language and often contain typos, non-standard acronyms, and spam. In addition to the textual content, Twitter data form a heterogeneous information network where users, tweets, and hashtags have mutual relationships. These features pose technical challenges for designing event detection and forecasting methods. In this talk, I will present the design and implementation of EMBERS, a fully automated 24x7 forecasting system for significant societal events using open source data including tweets, blog posts, and news articles. I will describe the system architecture of EMBERS, individual models that leverage specific data sources, and a fusion engine that supports trading off specific evaluation criteria. I will also demonstrate the superiority of EMBERS over base rate methods and its capability to forecast significant societal happenings.

ORGANIZERS
• Erik Cambria, Nanyang Technological University (Singapore)
• Bing Liu, University of Illinois at Chicago (USA)
• Yunqing Xia, Microsoft Research Asia (China)
• Yongzheng Zhang, LinkedIn Inc. (USA)

sentire2015


SENTIRE'14 (ICDM 2014, December 14th, Shenzhen)


PROGRAM
08:30-08:45 Welcoming and introduction (Erik Cambria)
08:45-09:30 From Big Data to Smart Nation: The Social Media Mining Perspective (Ee-Peng Lim)
09:30-10:00 A Semi-supervised Self-Adaptive Classifier over Opinionated Streams (Max Zimmermann)

10:00-10:15 Coffee break

10:30-11:00 Social Preference Ontologies for Enriching User and Item Data in Recommendation Systems (Christopher Krauss)
11:00-11:30 Joint Propagation and Refinement for Mining Opinion Words and Targets (Qiyun Zhao)
11:30-12:00 Semi-Supervised Method for Multi-Category Emotion Recognition in Tweets (Valentina Sintsova)
12:00-12:30 A Hybrid Approach for Emotion Detection in Support of Affective Interaction (Sonja Gievska)

12:30-14:00 Lunch break

14:00-14:30 A Localization Toolkit for SenticNet (Xiaoyu Li)
14:30-15:00 Towards Summarizing Popular Information from Massive Tourism Blogs (Hua Yuan)
15:00-15:30 Emotion Recognition from Text Based on Automatically Generated Rules (Wassim El-Hajj)

15:45-16:00 Coffee break

16:00-16:30 Commonsense Knowledge as the Glue in a Hybrid Model of Computational Creativity (Andrés Gómez)
16:30-17:00 Cyberbullying Detection using Time Series Modeling (Nektaria Potha)

SPEAKER
Ee-Peng Lim is a Professor at the School of Information Systems of Singapore Management University (SMU). He received Ph.D. from the University of Minnesota, Minneapolis in 1994 and B.Sc. in Computer Science from National University of Singapore. His research interests include social network and web mining, information integration, and digital libraries. He is currently an Associate Editor of the ACM Transactions on Information Systems (TOIS), Information Processing and Management (IPM), Social Network Analysis and Mining, Journal of Web Engineering (JWE), IEEE Intelligent Systems, International Journal of Digital Libraries (IJDL) and International Journal of Data Warehousing and Mining (IJDWM). He was a member of the ACM Publications Board until December 2012. He serves on the Steering Committee of the International Conference on Asian Digital Libraries (ICADL), Pacific Asia Conference on Knowledge Discovery and Data Mining (PAKDD), and International Conference on Social Informatics (SocInfo).

KEYNOTE
Social media mining research has been extremely vibrant in the last few years as much social media data are available to researchers for data science works. The relevant research topics covered span from topic trend and event analysis to social community and event discovery. In all these topics, sentiment and emotion expressed in social media are important components of the modeling and empirical research. In this talk, we will give an overview of the use of sentiment and emotion analyses to extract insights about users, communities, businesses, and events. Such insights can benefit many social and business applications including recommendations, customer relationship management, and policy making. We will also showcase a realtime social media analytics system that incorporates sentiment and emotion analysis.

ORGANIZERS
• Erik Cambria, Nanyang Technological University (Singapore)
• Bing Liu, University of Illinois at Chicago (USA)
• Yunqing Xia, Tsinghua University (China)
• Yongzheng Zhang, LinkedIn Inc. (USA)


sentire2014


SENTIRE'13 (ICDM 2013, December 7th, Dallas)


PROGRAM
08:30-08:45 Welcoming and introduction (Erik Cambria)
08:45-09:30 A cross-corpus study of subjectivity identification using unsupervised learning (Yang Liu)
09:30-10:00 Multi-class sentiment analysis with clustering and score representation (Mohsen Farhadloo)

10:00-10:30 Coffee break

10:30-11:00 Interest analysis using semantic PageRank and social interaction content (Chung-chi Huang)
11:00-11:30 Learning the roles of directional expressions and domain concepts in financial news analysis (Pyry Takala)
11:30-12:00 Robust language learning via efficient budgeted online algorithms (Giuseppe Castellucci)

12:00-13:30 Lunch break

13:30-13:45 Possible usage of sentiment analysis for calculating vectors of felific calculus (Rafal Rzepka)
13:45-14:00 Interpreting or describing? Verb abstraction in the linguistic category model (Aleksander Wawer)
14:15-14:30 Sentiment analysis in news articles using sentic computing (Prashant Raina)
14:30-14:45 Enhancing sentiment classification performance using bi-tagged phrases (Basant Agarwal)
14:45-15:00 A framework of review analysis for enhancement of business decision making (Atika Qazi)

15:00-15:30 Coffee break

15:30-16:00 Joint and pipeline probabilistic models for fine-grained sentiment analysis: Extracting aspects, subjective phrases and their relations (Roman Klinger)
16:00-16:30 Dynamic construction of dictionaries for sentiment classification (Hanen Ameur)
16:30-17:00 Subjective Bayes method for word semantic similarity measurement (Junhua Wang)
17:00-17:30 Pattern enhanced topic models for information filtering (Yue Xu)
17:30-18:00 Concluding remarks (Erik Cambria)

SPEAKER
Yang Liu is currently an Associate Professor in the Computer Science Department at the University of Texas at Dallas (UTD). She received her B.S. and M.S degree from Tsinghua University, and Ph.D from Purdue University. She was a researcher at the International Computer Science Institute (ICSI) at Berkeley before she joined UTD in 2005. Dr. Liu's research interest is in speech and natural language processing. She has published over 100 papers in this field. Dr. Liu received the NSF CAREER award in 2009 and the Air Force Young Investigator Program (YIP) award in 2010. She is currently an Associate editor of IEEE Transactions on Audio, Speech, and Language Processing; ACM Transactions on Asian Language Information Processing; and Speech Communication.

KEYNOTE
Sentiment/opinion analysis has received increasing attention in the past decade. The dominant approach has been based on supervised machine learning methods. However, these methods require a reasonable amount of annotated training data and often suffer from mismatched training and testing conditions. In this talk, the keynote speaker will present a study using unsupervised/semi-supervised generative learning methods for subjectivity detection across different domains. An initial training set using simple lexicon information is created, and then two iterative learning methods with a base naive Bayes classifier to learn from unannotated data are evaluated. The first method is self-training, which adds instances with high confidence into the training set in each iteration. The second is a calibrated EM (expectation-maximization) method where we calibrate the posterior probabilities from EM such that the class distribution is similar to that in the real data. Both approaches are evaluated on three different domains: movie data, news, and meeting dialogues. The keynote speaker will discuss various findings about the impacting factors for the model behaviors and show the inherent differences across domains.

ORGANIZERS
• Erik Cambria, National University of Singapore (Singapore)
• Bing Liu, University of Illinois at Chicago (USA)
• Yunqing Xia, Tsinghua University (China)
• Ping Chen, University of Houston-Downtown (USA)

sentire2013


SENTIRE'12 (ICDM 2012, December 10th, Brussels)


PROGRAM
09:00-09:10 Welcoming and introduction (Erik Cambria)
09:10-10:00 Multimodal sentiment analysis (Rada Mihalcea)

10:00-10:30 Coffee break

10:30-11:00 How much supervision? Corpus-based lexeme sentiment estimation (Aleksander Wawer)
11:00-11:30 Domain adaptation using domain similarity- and domain complexity-based instance selection for cross-domain sentiment analysis (Robert Remus)
11:30-12:00 Sentiment polarity classification using statistical data compression models (Dominique Ziegelmayer)

12:00-13:30 Lunch break

13:30-14:00 Representing and resolving negation for sentiment analysis (Emanuele Lapponi)
14:00-14:30 Fine-grained product features extraction and categorization in reviews opinion mining (Sheng Huang)
14:30-15:00 Subjectivity-based features for sentiment classification: A study on two lexicons (Rahim Dehkharghani)
15:00-15:30 Learning domain-specific polarity lexicons (Gulsen Demiroz)

15:30-16:00 Coffee break

16:00-16:30 A regularized recommendation algorithm with probabilistic sentiment-ratings (Filipa Peleja)
16:30-17:00 Enriching SenticNet polarity scores through semi-supervised fuzzy clustering (Soujanya Poria)
17:00-17:30 Full spectrum opinion mining: Integrating domain, syntactic and lexical knowledge (Daniel Olsher)
17:30-18:00 Concluding remarks (Erik Cambria)

SPEAKER
Rada Mihalcea is an Associate Professor in the Department of Computer Science and Engineering at the University of North Texas. Her research interests are in computational linguistics, with a focus on lexical semantics, graph-based algorithms for natural language processing, and multilingual natural language processing. She serves or has served on the editorial boards of the Journals of Computational Linguistics, Language Resources and Evaluations, Natural Language Engineering, Research in Language in Computation, IEEE Transactions on Affective Computing, and Transactions of the Association for Computational Linguistics. She was a program co-chair for the Conference of the Association for Computational Linguistics (2011), and the Conference on Empirical Methods in Natural Language Processing (2009). She is the recipient of a National Science Foundation CAREER award (2008) and a Presidential Early Career Award for Scientists and Engineers (2009).

KEYNOTE
With more than 10,000 new videos posted online every day on social websites such as YouTube and Facebook, the internet is becoming an almost infinite source of information. One crucial challenge for the coming decade is to be able to harvest relevant information from this constant flow of multi-modal data. In this talk, the keynote speaker will introduce the task of multi-modal sentiment analysis, and present a method that integrates linguistic, audio, and visual features for the purpose of identifying sentiment in online videos. The invited speaker will first describe a novel dataset consisting of videos collected from the social media website YouTube and annotated for sentiment polarity. She will then show, through comparative experiments, that the joint use of visual, audio, and textual features greatly improves over the use of only one modality at a time. Finally, by running evaluations on datasets in English and Spanish, the keynote speaker will show that the method is portable and works equally well when applied to different languages.

ORGANIZERS
• Erik Cambria, National University of Singapore (Singapore)
• Bing Liu, University of Illinois at Chicago (USA)
• Yunqing Xia, Tsinghua University (China)
• Catherine Havasi, MIT Media Laboratory (USA)

sentire2012


SENTIRE'11 (ICDM 2011, December 11th, Vancouver)


PROGRAM
08:30-08:40 Welcoming and introduction (Erik Cambria)
08:40-09:30 Sentiment analysis: A discovery challenge (Bing Liu)
09:35-10:00 STARLET: Multi-document summarization of service and product reviews with balanced rating distributions (Giuseppe Di Fabbrizio)

10:00-10:30 Coffee break

10:30-10:55 Multilingual sentiment analysis using latent semantic indexing and machine learning (Philip Kegelmeyer)
11:00-11:25 Deriving insights from national happiness indices (Daniel Archambault)
11:30-11:55 Multi-aspect sentiment analysis with topic models (Myle Ott)
12:00-12:30 AQA: Aspect-based opinion question answering (Samaneh Moghaddam)

12:30-14:00 Lunch break

14:00-14:15 A method for improving sentiment classification using feature highlighting (Lin Dai)
14:20-14:35 Machine reading for notion-based sentiment mining (Roula Hobeica)
14:40-14:55 Mining opinion attributes from texts using multiple kernel learning (Aleksander Wawer)
15:00-15:15 Longitudinal sales responses with online reviews (Wenyin Liu)
15:20-15:35 Fine-grained opinion mining using conditional random fields (Shabnam Shariaty)
15:40-16:00 Discourse structure and sentiment (Livia Polanyi)

16:00-16:30 Coffee break

16:30-16:55 Detecting general opinions from customer surveys (Evgeny Stepanov)
17:00-17:25 SES: Sentiment elicitation system for social media data (Yusheng Xie)
17:30-18:00 Concluding remarks (Erik Cambria)

SPEAKER
Bing Liu is a professor of Computer Science at University of Illinois at Chicago (UIC). He received his PhD in Artificial Intelligence from the University of Edinburgh. Before joining UIC, he was with the National University of Singapore. His current research interests include opinion mining and sentiment analysis, Web mining, and data mining. He has published extensively in leading conferences and journals in these fields. He has also written a textbook titled “Web Data Mining: Exploring Hyperlinks, Contents and Usage Data” published by Springer. The second edition of the book came out in July 2011. On professional services, Liu has served as program chairs of ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), IEEE International Conference on Data Mining (ICDM), ACM Conference on Web Search and Data Mining (WSDM), SIAM Conference on Data Mining (SDM), ACM Conference on Information and Knowledge Management (CIKM), and Pacific Asia Conference on Data Mining (PAKDD). Additionally, he has also served as associate editors of IEEE Transactions on Knowledge and Data Engineering (TKDE), Journal of Data Mining and Knowledge Discovery (DMKD), and SIGKDD Explorations, and is on the editorial boards of several other journals.

KEYNOTE
Sentiment analysis and opinion mining is the computational study of people’s opinions, evaluations, appraisals, attitudes, and emotions toward entities, issues, events, topics and their attributes. It has become a very active research area in natural language processing (NLP) and text mining due to many challenging research problems and a wide arrange of applications. The research has also spread from computer science to management science and social sciences. Although a large number of papers have been published, the pace of the progress has not been very fast due to the difficulty of natural language understanding. Many (if not most) existing papers do not make a lot of practical impact. One of the bottlenecks to the progress is the lack of effective mining algorithms that can discover complex expressions and domain commonsense knowledge which are essential for determining sentiment orientations and for recognizing targets of opinions. In this talk, Professor Bing Liu will first introduce the sentiment analysis and opinion mining problem, and then discuss several sub-problems which need novel data mining algorithms that can work collaboratively with NLP techniques for their solution.

ORGANIZERS
• Erik Cambria, National University of Singapore (Singapore)
• Yangqiu Song, Microsoft Research Asia (China)
• Catherine Havasi, MIT Media Laboratory (USA)
• Amir Hussain, University of Stirling (UK)

sentire2011