We consider the task of automatically phrasing and computing top-k rankings over the information contained in common knowledge bases (KBs), such as YAGO or DBPedia. We assemble the thematic focus and ranking criteria of rankings by inspecting the present Subject, Predicate, Object (SPO) triples. Making use of numerical attributes contained in the KB we are also able to compute the actual ranking content, i.e., entities and their performances. We further discuss the integration of existing rankings into the ranking generation process for increased coverage and ranking quality. We report on first results obtained using the YAGO knowledge base.

In this work, a benchmark to evaluate the retrieval performance of soundtrack recommendation systems is proposed. Such systems aim at finding songs that are played as background music for a given set of images. The proposed benchmark is based on preference judgments, where relevance is considered a continuous ordinal variable and judgments are collected for pairs of songs with respect to a query (i.e., set of images). To capture a wide variety of songs and images, we use a large space of possible music genres, different emotions expressed through music, and various query-image themes. The benchmark consists of two types of relevance assessments: (i) judgments obtained from a user study, that serve as a ``gold standard'' for (ii) relevance judgments gathered through Amazon's Mechanical Turk. We report on the performance of two state-of-the-art soundtrack recommendation systems using the proposed benchmark.

Focusing on the top-K items according to a ranking criterion constitutes an important functionality in many different query answering scenarios. The idea is to read only the necessary information---mostly from secondary storage---with the ultimate goal to achieve low latency. In this work, we consider processing such top-K queries under the constraint that the result items are members of a specific set, which is provided at query time. We call this restriction a set-defined selection criterion. Set-defined selections drastically influence the pros and cons of an id-ordered index vs. a score-ordered index. We present a mathematical model that allows to decide at runtime which index to choose, leading to a combined index. To improve the latency around the break even point of the two indices, we show how to benefit from a partitioned score-ordered index and present an algorithm to create such partitions based on analyzing query logs. Further performance gains can be enjoyed using approximate top-K results, with tunable result quality. The presented approaches are evaluated using both real-world and synthetic data.

We demonstrate PICASSO, a novel approach to soundtrack recommendation. Given text, video, or image documents, PICASSO selects the best fitting music pieces, out of a given set of files, for instance, a user's personal mp3 collection. This task, commonly referred to as soundtrack suggestion, is non-trivial as it requires a lot of human attention and a good deal of experience, with master pieces distinguished, e.g., with the Academy Award for Best Original Score. We put forward PICASSO to solve this task in a fully automated way. We address the problem by extracting the required information, in form of music/screenshot samples, from available contemporary movies, making the training set easily obtainable. The training set is further extended with information acquired from movie scripts and subtitles, giving us a richer description of the action and atmosphere expressed in a particular movie scene. Although the number of applications for this approach is very large, we focus on two selected applications. First, we consider recommendation of the soundtrack for the slide show generation based on the given set of images. Second, we consider recommending a soundtrack as the background music for given audio books.

We study the problem of automatically assigning appropriate music pieces to a picture or, in general, series of pictures. This task, commonly referred to as soundtrack suggestion, is non-trivial as it requires a lot of human attention and a good deal of experience, with master pieces distinguished, e.g., with the Academy Award for Best Original Score. We put forward PICASSO to solve this task in a fully automated way. PICASSO makes use of genuine samples obtained from first-class contemporary movies. Hence, the training set can be arbitrarily large and is also inexpensive to obtain but still provides an excellent source of information. At query time, PICASSO employs a three-level algorithm. First, it selects for a given query image a ranking of the most similar screenshots taken, and subsequently, selects for each screenshot the most similar songs to the music played in the movie when the screenshot was taken. Last, it issues a top-K aggregation algorithm to find the overall best suitable songs available. We have created a large training set consisting of over 40,000 image/soundtrack samples obtained from 28 movies and evaluated the suitability of PICASSO by means of a user study.