Apologies for cross-posting,
Participants are invited to the WWW’14 tutorial on Concept-Level Sentiment
Analysis, which will be held within the World Wide Web conference this April in
Seoul, Korea. The tutorial aims to provide its participants means to
efficiently design models, techniques, tools, and services for concept-level
sentiment analysis and their commercial realizations. The tutorial draws on
insights resulting from the recent IEEE Intelligent Systems special issues on
Concept-Level Opinion and Sentiment Analysis and the IEEE CIM special issue on
Computational Intelligence for Natural Language Processing. The tutorial
includes a hands-on session to illustrate how to build a concept-level
opinion-mining engine step-by-step, from semantic parsing to concept-level
reasoning.
BACKGROUND AND MOTIVATIONS
As the Web rapidly evolves, Web users are evolving with it. In an era of social
connectedness, people are becoming increasingly enthusiastic about interacting,
sharing, and collaborating through social networks, online communities, blogs,
Wikis, and other online collaborative media. In recent years, this collective
intelligence has spread to many different areas, with particular focus on
fields related to everyday life such as commerce, tourism, education, and
health, causing the size of the Social Web to expand exponentially.
The distillation of knowledge from such a large amount of unstructured
information, however, is an extremely difficult task, as the contents of
today’s Web are perfectly suitable for human consumption, but remain hardly
accessible to machines. The opportunity to capture the opinions of the general
public about social events, political movements, company strategies, marketing
campaigns, and product preferences has raised growing interest both within the
scientific community, leading to many exciting open challenges, as well as in
the business world, due to the remarkable benefits to be had from marketing and
financial market prediction.
Mining opinions and sentiments from natural language, however, is an extremely
difficult task as it involves a deep understanding of most of the explicit and
implicit, regular and irregular, syntactical and semantic rules proper of a
language. Existing approaches mainly rely on parts of text in which opinions
and sentiments are explicitly expressed such as polarity terms, affect words
and their co-occurrence frequencies. However, opinions and sentiments are often
conveyed implicitly through latent semantics, which make purely syntactical
approaches ineffective.
Concept-level sentiment analysis focuses on a semantic analysis of text through
the use of web ontologies or semantic networks, which allow the aggregation of
conceptual and affective information associated with natural language opinions.
By relying on external knowledge, such approaches step away from blind use of
keywords and word co-occurrence count, but rather rely on the implicit features
associated with natural language concepts. Unlike purely syntactical
techniques, concept-based approaches are able to detect also sentiments that
are expressed in a subtle manner, e.g., through the analysis of concepts that
do not explicitly convey any emotion, but which are implicitly linked to other
concepts that do so. The bag-of-concepts model can represent semantics
associated with natural language much better than bags-of-words. In the
bag-of-words model, in fact, a concept such as cloud computing would be split
into two separate words, disrupting the semantics of the input sentence (in
which, for example, the word cloud could wrongly activate concepts related to
weather).
The analysis at concept-level allows for the inference of semantic and
affective information associated with natural language text and, hence, enables
comparative fine-grained feature-based sentiment analysis. Rather than
gathering isolated opinions about a whole item (e.g., iPhone5), users are
generally more interested in comparing different products according to specific
features (e.g., iPhone5’s vs Galaxy S3’s touchscreen), or even sub-features
(e.g., fragility of iPhone5’s vs Galaxy S3’s touchscreen). In this context, the
construction of comprehensive common and common-sense knowledge bases is key
for feature-spotting and polarity detection, respectively. Common-sense, in
particular, is necessary to properly deconstruct natural language text into
sentiments – for example, to appraise the concept small room as negative for a
hotel review and small queue as positive for a post office, or the concept go
read the book as positive for a book review but negative for a movie review.
TUTORIAL PROGRAM
• Introduction (5 mins)
• New Avenues in Sentiment Analysis Research
- From Heuristics to Discourse Structure (5 mins)
- From Coarse to Fine-Grained Analysis (5 mins)
- From Keywords to Concepts (10 mins)
• Concept-Level Models
- Knowledge acquisition models (10 mins)
- Emotion categorization models (10 mins)
- Vector space models (10 mins)
• Concept-Level Techniques
- Analogical reasoning (10 mins)
- Parallel analogy (10 mins)
- Spreading activation (10 mins)
• Concept-Level Tools
- Sentiment resources (15 mins)
- Common knowledge repositories (15 mins)
- Aspect mining and polarity detection (10 mins)
• Building a Concept-Level Opinion-Mining Engine
- Semantic parsing (15 mins)
- Sentic API (15 mins)
- Application Samples (20 mins)
• Conclusion (5 mins)
IMPACT AND RELEVANCE
The World Wide Web Conference is a global event bringing together key
researchers, innovators, decision-makers, technologists, and business experts
trying to make meaning out of Web data. Within this research and business area,
opinion mining and sentiment analysis have become increasingly important
subtasks in recent years. However, there are still many challenges, including
social information understanding and integration, that need to be addressed.
For these reasons, a tutorial on concept-level sentiment analysis is strongly
relevant to WWW’14.
TARGET AUDIENCE AND PREREQUISITES
The target audience includes researchers and professionals in the fields of
sentiment analysis, Web data mining, and related areas. The tutorial also aims
to attract researchers from industry community as it covers research efforts
for the development of applications in fields such as commerce, tourism,
education, and health. The audience is expected to have basic computer science
skills, but psychologists and sociologists are also very welcome. The tutorial
not only covers state-of-the-art approaches to concept-level sentiment
analysis, but also provides information about techniques and tools to be used
for practical opinion mining.
ABOUT THE TUTOR
Erik Cambria received his BEng and MEng with honors in Electronic Engineering
from the University of Genova, in 2005 and 2008 respectively. In 2011, he has
been awarded a PhD in Computing Science and Mathematics, following the
completion of an industrial Cooperative Awards in Science and Engineering
(CASE) research project, funded by the UK Engineering and Physical Sciences
Research Council (EPSRC), which was born from the collaboration between the
University of Stirling and the MIT Media Laboratory.
Today, Erik is the lead investigator of a MINDEF-funded project on Commonsense
Knowledge Representation & Reasoning at the National University of Singapore
(Temasek Laboratories) and an associate researcher at the MIT Media Laboratory
(Synthetic Intelligence Project). His interests include AI, Semantic Web, KR,
NLP, opinion mining and sentiment analysis, affective and cognitive modeling,
intention awareness, HCI, and e-health. Erik is also chair of several
international conferences, e.g., Extreme Learning Machines (ELM), and workshop
series, e.g., ICDM SENTIRE. He is on the editorial board of Springer Cognitive
Computation and he is a guest editor of many other leading AI journals. Erik is
also a fellow of the Brain Sciences Foundation, the Chinese Academy of
Sciences, National Taiwan University, Microsoft Research Asia, and HP Labs
India.
