Paul Seitlinger

Recommending tags with a model of human categorization

When interacting with social tagging systems, humans exercise complex processes of categorization that have been the topic of much research in cognitive science. In this paper we present a recommender approach for social tags derived from ALCOVE, a model of human category learning. The basic architecture is a simple three-layers connectionist model. The input layer encodes patterns of semantic features of a user-specific resource, such as latent topics elicited through Latent Dirichlet Allocation (LDA) or available external categories. The hidden layer categorizes the resource by matching the encoded pattern against already learned exemplar patterns. The latter are composed of unique feature patterns and associated tag distributions. Finally, the output layer samples tags from the associated tag distributions to verbalize the preceding categorization process. We have evaluated this approach on a real-world folksonomy gathered from Wikipedia bookmarks in Delicious. In the experiment our approach outperformed LDA, a well-established algorithm. We attribute this to the fact that our approach processes semantic information (either latent topics or external categories) across the three different layers. With this paper, we demonstrate that a theoretically guided design of algorithms not only holds potential for improving existing recommendation mechanisms, but it also allows us to derive more generalizable insights about how human information interaction on the Web is determined by both semantic and verbal processes.

Modeling Activation Processes in Human Memory to Predict the Use of Tags in Social Bookmarking Systems

Modeling Activation Processes in Human Memory to Predict the Use of Tags in Social Bookmarking Systems

Implicit imitation in social tagging: familiarity and semantic reconstruction

Social Tagging is a recent widespread phenomenon on the Web where people assign labels (tags) to Web resources. It has been hypothesized to support collaborative sensemaking. In this paper, we examine some of the cognitive mechanisms assumed to underlie sensemaking, namely social imitation. In line with the semantic imitation model of Fu et al., we assume that implicit processing can be understood as a semantic reconstruction of gist. Our model contrasts this process with a recall of tags from an explicit verbatim memory trace. We tested this model in an experimental study in which after the search task students had to generate tags themselves. We exposed their answers to a multinomial model derived from Fuzzy Trace Theory to obtain independent parameter estimates for the processes of explicit recall, semantic gist reconstruction and familiarity-based recall. A model that assumes all processes are at play explains the data well. Similar to results of our previous study, we find an influence of search intentions on the two processes. Our results have implications for interface and interaction design of social tagging systems, as well as for tag recommendation in these environments.

Towards a user model for personalized recommendations in work-integrated learning: A report on an experimental study with a collaborative tagging system

Abstract The informal setting of learning at work give rise for unique challenges to the field of technology enhanced learning systems. Personalized recommendations taking into account the current context of the individual knowledge worker are a powerful approach to overcome those challenges and effectively support the knowledge workers to meet their individual information needs. Basis for these recommendations to adopt to the current context of a knowledge worker can be provided by user models which reflects the topics knowledge workers are dealing with and their corresponding knowledge levels, but research has only focused on user modeling in settings with a static underlying domain model so far. We suggest to model the users’ context based on the emergent topics they are dealing with and their individual current knowledge levels within these topics by extracting the necessary information from the user’s past activities within the system. Based on data from an experiment with students learning a new topic with the help of a collaborative tagging system, we started to evaluate this approach and report on first results.

Forgetting the Words but Remembering the Meaning: Modeling Forgetting in a Verbal and Semantic Tag Recommender

We assume that recommender systems are more successful, when they are based on a thorough understanding of how people process information. In the current paper we test this assumption in the context of social tagging systems. Cognitive research on how people assign tags has shown that they draw on two interconnected levels of knowledge in their memory: on a conceptual level of semantic fields or LDA topics, and on a lexical level that turns patterns on the semantic level into words. Another strand of tagging research reveals a strong impact of time-dependent forgetting on users’ tag choices, such that recently used tags have a higher probability being reused than “older” tags. In this paper, we align both strands by implementing a computational theory of human memory that integrates the two-level conception and the process of forgetting in form of a tag recommender. Furthermore, we test the approach in three large-scale social tagging datasets that are drawn from BibSonomy, CiteULike and Flickr.

Attention Please! A Hybrid Resource Recommender Mimicking Attention-Interpretation Dynamics

Classic resource recommenders like Collaborative Filtering (CF) treat users as being just another entity, neglecting non-linear user-resource dynamics shaping attention and interpretation. In this paper, we propose a novel hybrid recommendation strategy that refines CF by capturing these dynamics. The evaluation results reveal that our approach substantially improves CF and, depending on the dataset, successfully competes with a computationally much more expensive Matrix Factorization variant.

Dynamics of human categorization in a collaborative tagging system

Highlights • We study how categories people develop in collaborative tagging change over time.• Their internal cognitive categories and the tags they use are coordinated.• Especially groups converging in the use of terms develop differentiated categories.• Social processes around shared artefacts have a mediating effect on learning.

Refining Frequency-Based Tag Reuse Predictions by Means of Time and Semantic Context

In this paper, we introduce a tag recommendation algorithm that mimics the way humans draw on items in their long-term memory. Based on a theory of human memory, the approach estimates a tag’s probability being applied by a particular user as a function of usage frequency and recency of the tag in the user’s past. This probability is further refined by considering the influence of the current semantic context of the user’s tagging situation. Using three real-world folksonomies gathered from bookmarks in BibSonomy, CiteULike and Flickr, we show how refining frequency-based estimates by considering usage recency and contextual influence outperforms conventional “most popular tags” approaches and another existing and very effective but less theory-driven, time-dependent recommendation mechanism.

Evaluating the validity of a non-invasive assessment procedure

Recent developments in serious games allow for in-game adaptations to enhance the learners current cognitive, motivational or emotional state. Providing suitable adaptations requires a valid assessment of the psycho-pedagogical constructs the game should adapt to. An explicit assessment, e.g. by questionnaires occurring repeatedly on the screen, would impair the learners game flow. Therefore, a non-invasive and implicit assessment procedure is required. In the course of the European research project TARGET, we established an assessment procedure which is based on the interpretation of the learners actions in the virtual environment, called Behavioural Indicators (BIs). A set of 16 BIs has been formulated to assess the learners current emotional, motivational and clearness state. In this present work, we describe how these BIs can be validated and focus on the innovative elements of the methodological procedure, the material, experiential considerations and the statistical analysis to be applied in an empirical study.

Implicit and explicit memory in learning from social software: a dual-process account

Inspired by a recent surge to understand social cognitive processes in collaborative knowledge building, we have devised an experiment in which students learned from contents of a wiki. One of the informative results we observed was a dissociation between implicit and explicit memory measures that we used to track students learning: an association test, and the drawing of concept maps, respectively. We put these initial results in the context of experimental research in cognitive psychology and show how the co-evolution model (Cress and Kimmerle, 2008) could account for them. With several network measures, we also suggest some ways of how to measure assimilation and accommodation, both in internal and external knowledge representations.