Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining - Concepts and Techniques.

An important part of our information-gathering behavior has always been to find out what other people think. With the growing availability and popularity of opinion-rich resources such as online review sites and personal blogs, new opportunities and challenges arise as people now can, and do, actively use information technologies to seek out and understand the opinions of others. The sudden eruption of activity in the area of opinion mining and sentiment analysis, which deals with the computational treatment of opinion, sentiment, and subjectivity in text, has thus occurred at least in part as a direct response to the surge of interest in new systems that deal directly with opinions as a first-class object.

This survey covers techniques and approaches that promise to directly enable opinion-oriented information-seeking systems. Our focus is on methods that seek to address the new challenges raised by sentiment-aware applications, as compared to those that are already present in more traditional fact-based analysis. We include material on summarization of evaluative text and on broader issues regarding privacy, manipulation, and economic impact that the development of opinion-oriented information-access services gives rise to. To facilitate future work, a discussion of available resources, benchmark datasets, and evaluation campaigns is also provided.

/pdf/opinion-mining-and-sentiment-analysis-3v5ya3qqiw.pdf

Opinion Mining and Sentiment Analysis

From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

We present a system for interactively browsing and exploring large unstructured collections of photographs of a scene using a novel 3D interface. Our system consists of an image-based modeling front end that automatically computes the viewpoint of each photograph as well as a sparse 3D model of the scene and image to model correspondences. Our photo explorer uses image-based rendering techniques to smoothly transition between photographs, while also enabling full 3D navigation and exploration of the set of images and world geometry, along with auxiliary information such as overhead maps. Our system also makes it easy to construct photo tours of scenic or historic locations, and to annotate image details, which are automatically transferred to other relevant images. We demonstrate our system on several large personal photo collections as well as images gathered from Internet photo sharing sites.

Photo tourism: exploring photo collections in 3D

In next-generation networking environments, ubiquitous networks will be available both indoors and outdoors. Various devices will be ubiquitously embedded in our homes and cityscape. Digital content will be stored not only by servers on the Internet, but also in embedded devices belonging to ubiquitous networks. In this paper, we propose a content-processing mechanism for use in environment-enabling collaborative acquisition of embedded digital content in real-world situations. We have developed a network management device that makes it possible to acquire embedded content using coordinated ubiquitous devices. This management device actively configures networks that include content-providing devices and browsing devices to enable sharing of a range of digital content. To demonstrate our system, we built a practical prototype called the "Virtual Insect Catching System", which is simple enough for children to use. In a test that 48 children took part in we demonstrated that the system can be used to find embedded devices, build peer-topeer networks, acquire embedded digital content, retrieve related content from the Internet, and then create new web content.

Cooperative Device Browsing through Portable Private Area Network

In this paper, we propose a trust-oriented evaluation method for information retrieval methods based on computing the content coverage of the search results. At first, for a user given query, we collect the related resources available on the Internet. We then sketch out an overview (called topic-sketch) of these resources. We compute the content coverage of the search results to evaluate the retrieval methods to see whether they comprehensively find and provide us information, by comparing the resource overviews and each result. In other words, the content coverage is also useful for estimating the content comprehensibility provided by a Web page on a certain topic. Therefore, we can also use the content coverage as a measure to re-rank the search results. Some preliminary experimental results are shown in this paper to validate our method.

Content-coverage Based Trust-oriented Evaluation Method for Information Retrieval

The Web contains an immense variety of images. Because an image is typically published as part of some Web content, it needs to be retrieved based not only on its content, but also based on the relations between it and the Web contents surrounding it. That is, the Web contents represent the "context" of the image within the Web. A context semantically represents the usage and/or reputation of an image. The context also helps a user select reliable images from the Web, which is difficult when based on only the image content properties. In this paper, we propose a concept of "context" in the Web, and show how the context can be extracted and visualized using our prototypes.

Retrieving Web images based on their usage context for augmenting ubiquitous contents

Web3D technologies have made it possible for users to browse and manipulate 3D CG model data on the Web. Usually, such data requires users to interact with it in order to see its attribute information and its attached behavior. To automatically present 3D CG data and its attribute information, it is necessary to create 3D CG animation in advance. However, much time and cost is needed to create 3D CG animation that introduces its shape and its attribute information. We aim at constructing a mechanism that automatically produces 3D CG animation from the original 3D CG model data and its attribute information. 3D CG data has become popular on mobile phones such as i-mode. It is, however, difficult for users to interact with such data on mobile phones, because of their limited interaction capability and small-size screen. To cope with the problem, we also need to have a mechanism that produces 3D CG animation in an autonomous manner. We propose a way to automatically generate 3D CG animation from the original 3D CG model data and its attribute information. With the proposed method, users can browse 3D CG data and its attribute information in a "less clicking and more watching" manner. The proposed system automatically generates CG animation synchronized with synthesized speech from 3D CG contents and their attribute data. We also propose a method to generate a presentation for multiple 3D CG objects, that emphasizes the major differences in their attribute information.

Autonomous presentation of 3 dimensional CG contents on the Web

The authors propose the idea of associating and presenting metadata (attribute information) of digital archive objects in a virtual space. In the presentation using a virtual space, we introduce two methods. One is called InfoLOD, which dynamically associates and controls metadata (attribute information) with archive objects (2D/3D images) in the 3D space. We propose to control the level of detail (LOD) of attribute information based on orientation and distance. We extend this idea to control the LOD of hierarchically-described metadata (such as XML data) based on distance, orientation, and differentiation. LOD control based on differentiation compares the hierarchical data of spatial objects visible from one's viewpoint in 3D space, and presents the dissimilarities in the objects having similar data. The other method is called passive viewing, and it realizes users to watch and listen automatic presentation of 3D objects and their metadata. Some of our implementations are shown to illustrate our ideas.

Katsumi Tanaka

Papers

Cooperative Device Browsing through Portable Private Area Network

Content-coverage Based Trust-oriented Evaluation Method for Information Retrieval

Retrieving Web images based on their usage context for augmenting ubiquitous contents

Autonomous presentation of 3 dimensional CG contents on the Web

Associating and presenting metadata of digital archive objects in virtual spaces as synchronized multimedia content