Matrix Factorization Techniques for Recommender Systems

Cultivating Communities of Practice: A Guide to Managing Knowledge

Cognitive Surplus: Creativity and Generosity in a Connected Age. Clay Shirky. New York, NY: Penguin, 2010. 242 pages, $25.95 pbk. It sometimes seems that the hardest thing to do in the Information Age is to communicate. In the rush of easily accessible data and the maelstrom of conflicting viewpoints, two otherwise intelligent people can talk past one another as they stake out territory with the tenacity of computer viruses. NYU professor Clay Shirky and media critic Nicholas Carr have been squaring off now for two years over what impact the Internet is having on our society. Shirky takes the more optimistic viewpoint, Carr the more pessimistic. Carr threw down the gauntlet with his 2008 Atlantic cover article "Is Google Making Us Stoopid?" and has continued the debate with his recent book The Shallows. Shirky provides a response in Cognitive Surplus, but criticizes skeptics like Carr only obliquely. Shirky says their main frustration is with the profusion of choice: "Scarcity," he says, "is easier to deal with than abundance." But this book doesn't dwell much on the naysayers. Instead, it frames the subject of the Internet with a bold and startling vision about its potential. "Imagine," he says, "treating the free time of the world's educated citizenry as an aggregate, a kind of cognitive surplus." We've not always used that surplus wisely. Shirky's case in point is television, and how it has come to dominate our culture. Over much of the planet, Shirky writes, "the three most common activities are . . . work, sleep, and watching TV." Like gin in early-eighteenth-century London, twentieth-century television is one of those social habits that critics have denounced and tried hard to minimize, but without success. There are signs now, however, that TV viewing - still thoroughly popular - isn't quite the juggernaut it used to be. Young people are increasingly turning to the Internet; and the Web, it turns out, allows humans to do things they can't do with other media - namely, create, produce, and connect. Instead of devoting twenty passive hours a week to the tube (the international average), people now use a medium that lets them make and share things. That may seem trivial, considering the amount of silly, offensive, or deceptive fare on the Web, but think of it this way: Which do you think contains more enduring cultural, intellectual, and societal value - posting comments on a blog or watching Gitligan's Island! The Internet is no utopia, but neither are older media. In fact, some of them may be a good deal less salutary. Take, for example, the online fantasy game World of Warcraft. As Shirky puts it in a tart retort: "However pathetic you may think it is to sit in your basement pretending to be an elf, I can tell you from personal experience: it's worse to sit in your basement trying to decide whether Ginger or Mary Ann is cuter." There's a wide variety of content on the Internet, of course - the widest in any medium - and the largest number of producers of content: a colossal and revolutionary force. The "hundred million hours of cumulative thought" it took to produce Wikipedia is one example of Shirky's "cognitive surplus," and while it still pales before the "two hundred billion hours of TV every year," you can readily see the potential. …

Cognitive Surplus: Creativity and Generosity in a Connected Age

The increasing availability of (digital) cultural heritage artefacts offers great potential for increased access to art content, but also necessitates tools to help users deal with such abundance of information. User-adaptive art recommender systems aim to present their users with art content tailored to their interests. These systems try to adapt to the user based on feedback from the user on which artworks he or she finds interesting. Users need to be able to depend on the system to competently adapt to their feedback and find the artworks that are most interesting to them. This paper investigates the influence of transparency on user trust in and acceptance of content-based recommender systems. A between-subject experiment (N = 60) evaluated interaction with three versions of a content-based art recommender in the cultural heritage domain. This recommender system provides users with artworks that are of interest to them, based on their ratings of other artworks. Version 1 was not transparent, version 2 explained to the user why a recommendation had been made and version 3 showed a rating of how certain the system was that a recommendation would be of interest to the user. Results show that explaining to the user why a recommendation was made increased acceptance of the recommendations. Trust in the system itself was not improved by transparency. Showing how certain the system was of a recommendation did not influence trust and acceptance. A number of guidelines for design of recommender systems in the cultural heritage domain have been derived from the study's results.

/pdf/the-effects-of-transparency-on-trust-in-and-acceptance-of-a-t5rtj3cgc6.pdf

The effects of transparency on trust in and acceptance of a content-based art recommender

Graph Convolutional Networks (GCNs) have recently been shown to be quite successful in modeling graph-structured data. However, the primary focus has been on handling simple undirected graphs. Multi-relational graphs are a more general and prevalent form of graphs where each edge has a label and direction associated with it. Most of the existing approaches to handle such graphs suffer from over-parameterization and are restricted to learning representations of nodes only. In this paper, we propose CompGCN, a novel Graph Convolutional framework which jointly embeds both nodes and relations in a relational graph. CompGCN leverages a variety of entity-relation composition operations from Knowledge Graph Embedding techniques and scales with the number of relations. It also generalizes several of the existing multi-relational GCN methods. We evaluate our proposed method on multiple tasks such as node classification, link prediction, and graph classification, and achieve demonstrably superior results. We make the source code of CompGCN available to foster reproducible research.

/pdf/composition-based-multi-relational-graph-convolutional-4ntuyr3sol.pdf

Composition-based Multi-Relational Graph Convolutional Networks

https://www.cs.vu.nl/~guus/papers/Schreiber08a.pdf

Semantic annotation and search of cultural-heritage collections: The MultimediaN E-Culture demonstrator

The main objective of the MultimediaN E-Culture project is to demonstrate how novel semantic-web and presentation technologies can be deployed to provide better indexing and search support within large virtual collections of cultural-heritage resources. The architecture is fully based on open web standards, in particular XML, SVG, RDF/OWL and SPARQL. One basic hypothesis underlying this work is that the use of explicit background knowledge in the form of ontologies/vocabularies/thesauri is in particular useful in information retrieval in knowledge-rich domains.

/pdf/multimedian-e-culture-demonstrator-4393awrau0.pdf

MultimediaN e-culture demonstrator

Within the cultural heritage field, proprietary metadata and vocabularies are being transformed into public Linked Data. These efforts have mostly been at the level of large-scale aggregators such as Europeana where the original data is abstracted to a common format and schema. Although this approach ensures a level of consistency and interoperability, the richness of the original data is lost in the process. In this paper, we present a transparent and interactive methodology for ingesting, converting and linking cultural heritage metadata into Linked Data. The methodology is designed to maintain the richness and detail of the original metadata. We introduce the XMLRDF conversion tool and describe how it is integrated in the ClioPatria semantic web toolkit. The methodology and the tools have been validated by converting the Amsterdam Museum metadata to a Linked Data version. In this way, the Amsterdam Museum became the first ‘small' cultural heritage institution with a node in the Linked Data cloud.

http://www.few.vu.nl/~vbr240/publications/eswc2012supporting.pdf

Supporting linked data production for cultural heritage institutes: the amsterdam museum case study

In modern machine learning, raw data is the pre-ferred input for our models. Where a decade ago data scien-tists were still engineering features, manually picking out the details they thought salient, they now prefer the data in their raw form. As long as we can assume that all relevant and ir-relevant information is present in the input data, we can de-sign deep models that build up intermediate representations to sift out relevant features. However, these models are often domain specific and tailored to the task at hand, and therefore unsuited for learning on heterogeneous knowledge: informa-tion of different types and from different domains. If we can develop methods that operate on this form of knowledge, we can dispense with a great deal of ad-hoc feature engineering and train deep models end-to-end in many more domains. To accomplish this, we first need a data model capable of ex-pressing heterogeneous knowledge naturally in various do-mains, in as usable a form as possible, and satisfying as many use cases as possible. In this position paper, we argue that the knowledge graph is a suitable candidate for this data model. This paper describes current research and discusses some of the promises and challenges of this approach.

/pdf/the-knowledge-graph-as-the-default-data-model-for-learning-2vy6v0ewp2.pdf

The knowledge graph as the default data model for learning on heterogeneous knowledge

In this position paper we identify nichesourcing, a specific form of human-based computation that harnesses the computational efforts from niche groups rather than the "faceless crowd". We claim that nichesourcing combine the strengths of the crowd with those of professionals, optimizing the result of human-based computation for certain tasks. We illustrate our claim using scenarios in two domains: cultural heritage and regreening in Africa. The contribution of this paper is to provide a definition of the main characteristics of nichesourcing as a natural extension of crowdsourcing and to outline research challenges for realizing nichesourcing applications.

/pdf/nichesourcing-harnessing-the-power-of-crowds-of-experts-14yg7q8mum.pdf

Victor de Boer

Papers

Semantic annotation and search of cultural-heritage collections: The MultimediaN E-Culture demonstrator

MultimediaN e-culture demonstrator

Supporting linked data production for cultural heritage institutes: the amsterdam museum case study

The knowledge graph as the default data model for learning on heterogeneous knowledge

Nichesourcing: harnessing the power of crowds of experts