Search

Research into recommendation algorithms has made great strides in recent years. However, these algorithms are typically applied in relatively straightforward scenarios: given information about a user's past preferences, what will they like in the future? Recommendation is often more complex: evaluating recommended items never takes place in a vacuum, and it is often a single step in the user's more complex background task. In this paper, we define a specific type of recommendation scenario called narrative-driven recommendation, where the recommendation process is driven by both a log of the user's past transactions as well as a narrative description of their current interest(s). Through an analysis of a set of real-world recommendation narratives from the LibraryThing forums, we demonstrate the uniqueness and richness of this scenario and highlight common patterns and properties of such narratives.

In this paper, we propose a novel framework using the word2vec model, a deep learning method, integrated with a book ontology in order to enhance semantically searching books. The idea starts from constructing a book ontology for reasoning book information efficiently. A deep learning method, namely the word2vec model, is then utilized to represent vectors of words occurring on book descriptions. These vectors would help finding most relevant books given a query string. The integration of the word2vec model and the book ontology is able to achieve high performance in searching books. A database of Amazon books is taken into account examining the proposed method, compared with an advanced keyword matching method. The experimental results show that the proposed method can produce more accurate searching results.

One way to reduce opportunity loss and help stop the proliferation of bad book reviews is to reduce the gap between expectation before reading and satisfaction after reading. A previously developed method for revising book cover images to reflect reader impressions reduces the gap between the impression obtained from a book cover image and the impression after reading the book. A book search system based on this method has now been constructed that enables users to easily search for books on the basis of impressions after reading as reflected by the revised book cover images. Its ring-structured graphical search interface supports multi-faceted search using simple ring operations. Testing showed that it enables users to easily search for books on the basis of impressions after reading and enhances the opportunity to discover new books during the search process.

We thank our funding agencies—NSERC, SurfNet, AITF, Grand, Smart Technologies—that made this research possible. Today, searching through collections of books or written texts is often performed using digital search engines. There, the efficient yet limiting query paradigm is still the most dominant entry point. Previous work has characterized search processes in various contexts and describes search as integral and closely related to creative endeavours. We revisit this work from a design perspective, proposing guidelines for versatile search interfaces that are based on a modular approach to search. Inspired by aspects of search in physical environments, our recommendations address learning, creativity, inspiration, and pleasure as important and positive factors of (book) search. Based on in-depth interviews with library patrons about search practises in physical and digital environments and drawing from previous work on search behaviour, we discuss search patterns as modular constructs consisting of micro-strategies. We illustrate how the structure of these patterns, much like creative processes, is highly flexible and fluidly evolves based on learning and ideation during search, particularly in physical search environments. This modular approach to search provides a concrete yet flexible basis to design interfaces that facilitate versatile and creative approaches to search in digital environments. Postprint

Multi-faceted book search engine presents diverse category-style options to allow users to refine search results without re-entering a query. In this paper, we propose a novel multi-faceted book search engine that utilizes users' query-related latent intents mined from click-through logs as multiple facets for books. The latent query intents can be effectively and efficiently discovered by applying the Sparse Latent Semantic Analysis (LSA) model to users' query and clicking behaviors in the click-through logs. This paper presents the details to improve the multi-faceted book search by incorporating the compact representation of query-intent-book relationships generated by Sparse LSA into the off-line and online processing procedures. The specificity of latent query intents can be flexibly changed by adjusting the sparsity level of projection matrix in the Sparse LSA model. We evaluated our approach on CADAL click-through logs containing 45,892 queries and 164,822 books. The experimental results show the Sparse LSA model with more sparse projection matrix tends to discover the more specific latent query intents. The latent query intents suggested by our approach usually gain the high user satisfaction ratio.

With massive book digitization efforts underway, there is a need for developing effective book retrieval strategies. This paper explores the relative contribution of different parts of digitized and OCR'ed books towards effective retrieval. The examined parts include the entire content of books, book headings, book titles, and table of content entries. Results show that indexing the headers and titles of books is nearly as effective as indexing the entire contents of books. These results indicate that certain portions of the books, specifically titles and headers, are more valuable than other parts of books. This is akin to web search where hypertext and page titles are more valuable to index than the rest of the webpage. Also, using a combination of evidence approach provides further improved retrieval effectiveness compared to using any portion of the book in isolation.

This paper describes the Active Reading task, a recent addition to the INEX Book Search track1. This task aims at exploring how people interact with e-books in different scenarios of use. Besides, it has been designed to take into account a corpus of research on e-book usability and make sure this does not become obsolete but actually inform the design of better, and more usable, e-books and e-readers in the future. This paper starts introducing the relevant research context in order to justify the crucial role of the Active Reading task. It then describes the rational behind the task and its origins. It goes on to examine its aims and objectives, by exploring its related research questions and looking at expectations it should raise, and the positive impact such initiative could have on the e-book research community.