Introduction - Methods for Mining and Summarizing Text Conversations - page 2

what opinions the conversation participants have expressed on such topics. We will also discuss how
the underlying structure of a text conversation can be determined by identifying specific dialogue
acts (e.g., request, answer) and their relationship (e.g., a question/answer pair). All these kinds of
extracted information, expressed in term of topics, opinions and conversational structure, can then
be used to summarize the conversation. We will see how different summaries can be generated
at different levels of granularity depending, for instance, on the audience and specific information
needs.
Most of the studies we cover in this topic involve techniques to deal with text conversations in
a particular modality (e.g., extracting opinions from an email conversation). However, we will also
discuss more recent work which is increasingly tackling the challenges of mining and summarizing
conversations spanning multiple modalities (e.g., a transcript of a meeting that was followed up by
an email conversation).
It is widely accepted that we are in the midst of an epochal revolution in how people communi-
cate. We believe that researchers have just started to envision the plethora of opportunities presented
by the widespread availability of text conversations. We hope that our readers, after learning the var-
ious techniques described in this topic for mining and summarizing text conversations, will be able
to find new and creative ways to apply them, making (computer-mediated) human communication
ever more effective.
In the remainder of this introduction, we first present some data on the amazing pace at
which human conversations are moving from the spoken to the text form. Then, we discuss some
key application scenarios for conversation mining and summarization. After that, we will conclude
with an overview of the research space for the computational techniques we will explore in this topic.
1.1 THE RISE OF TEXT CONVERSATIONS
In this section we first consider the rise of internet technologies and subsequently discuss contem-
poraneous developments in speech technology, and describe how both have transformed the way we
communicate with one another.
1.1.1 THE INTERNET REVOLUTION
The rapid adoption rate of new, Web-based forms of communication, has beaten even the most
hyperbolic predictions. Email was born in 1971 and is considered the grandparent of all Web-based
social media. Today, it is used daily by billions of people all over the world in a seemingly unlimited
variety of communicative settings. We email our friends to organize a trip over the weekend, our
colleagues to discuss next year's budget, and a car dealer to bargain the price of a new car. As Baron
[ 2008 ], a leading expert in Computer Mediated Communication, put it: “Emails have style and
content as diverse as people using it”. Also, email is clearly a domesticated technology, i.e., it has
become a normal component of daily living [ Baron , 2008 ]. Although precise data are always difficult
to come by, a quick Web search tells us that in 2009 Yahoo Mail had a billion emails passing through
its servers every day (as reported by its CEO, Carol Bartz). According to more recent data on websites