Probabilistic topic modeling in multilingual settings: An overview of its methodology and applications

Probabilistic topic models are unsupervised generative models which model document content as a two-step generation process, that is, documents are observed as mixtures of latent concepts or topics, while topics are probability distributions over vocabulary words. Recently, a significant research effort has been invested into transferring the probabilistic topic modeling concept from monolingual to multilingual settings. Novel topic models have been designed to work with parallel and comparable texts. We define multilingual probabilistic topic modeling (MuPTM) and present the first full overview of the current research, methodology, advantages and limitations in MuPTM. As a representative example, we choose a natural extension of the omnipresent LDA model to multilingual settings called bilingual LDA (BiLDA). We provide a thorough overview of this representative multilingual model from its high-level modeling assumptions down to its mathematical foundations. We demonstrate how to use the data representation by means of output sets of (i) per-topic word distributions and (ii) per-document topic distributions coming from a multilingual probabilistic topic model in various real-life cross-lingual tasks involving different languages, without any external language pair dependent translation resource: (1) cross-lingual event-centered news clustering, (2) cross-lingual document classification, (3) cross-lingual semantic similarity, and (4) cross-lingual information retrieval. We also briefly review several other applications present in the relevant literature, and introduce and illustrate two related modeling concepts: topic smoothing and topic pruning. In summary, this article encompasses the current research in multilingual probabilistic topic modeling. By presenting a series of potential applications, we reveal the importance of the language-independent and language pair independent data representations by means of MuPTM. We provide clear directions for future research in the field by providing a systematic overview of how to link and transfer aspect knowledge across corpora written in different languages via the shared space of latent cross-lingual topics, that is, how to effectively employ learned per-topic word distributions and per-document topic distributions of any multilingual probabilistic topic model in various cross-lingual applications.     

The role of self- and social directed goals in a problem-based, collaborative learning context

Students’ learning goals demonstrate much stronger variety than traditional goal orientation models for classroom learning assume, especially when the educational context allows so. In this empirical study we will investigate the richness of students’ goal orientation in a collaborative learning context. We do so with the help of a goal setting framework that is based on a two-facet approach distinguishing multiple contents (performance, learning, well-being) and goal directions (varying degrees of self vs social direction). To investigate the role of different goal constellations, goal setting and learning performance data of first year students (n = 2,636) in a problem-based, collaborative learning program, and evaluation data of problem-based tutorial groups (#groups = 206) are combined into a multilevel model. Each tutorial group functions in two different educational settings: one directed at open-ended, group problems, the other at closed, individual problems. Educational context appears indeed to have a crucial role in the relationship between students’ goal setting at the one side, and students’ performance and group functioning on the other side.     

Philosophy,education, and the explosion of knowledge

In science, and in the teaching of science, knowinghow is as important as knowingthat. The value of theoretical knowledge is limited, especially so in view of the rapid increase in available information. Methodological analysis of concrete problems should play a major role in teaching. Also, students should learn to contextualize problems in disciplines other than their major area of interest. Generalists are badly needed for this, in teaching and in research. However, current forms of interdisciplinary integration must be de-emphasized.     

Students’ Ability to Connect Function Properties to Different Types of Elementary Functions: An Empirical Study on the Role of External Representations

Recent research on the phenomenon of improper proportional reasoning focused on students’ understanding of elementary functions and their external representations. So far, the role of basic function properties in students’ concept images of functions remained unclear. We add to this research line by investigating how accurate students are in connecting functions to their corresponding properties and how this accuracy depends on function types and representations. A large group of 10th graders evaluated for different function types, represented in either a graphical, a formulaic, or a tabular mode, the correctness of statements about their general properties and behavior. Results show that students succeeded rather well in making the right connections between properties and functions. Errors depended not only on the type of function for which the properties were evaluated but also on the kind of representation in which the function was presented. These results highlight the importance of function properties in students’ concept images of functions and suggest positive effects of making these properties explicit to students.     

Performance assessment in serious games: Compensating for the effects of randomness

This paper is about performance assessment in serious games. We conceive serious gaming as a process of player-lead decision taking. Starting from combinatorics and item-response theory we provide an analytical model that makes explicit to what extent observed player performances (decisions) are blurred by chance processes (guessing behaviors). We found large effects both theoretically and practically. In two existing serious games random guess scores were found to explain up to 41 % of total scores. Monte Carlo simulation of random game play confirmed the substantial impact of randomness on performance. For valid performance assessments, be it in-game or post-game, the effects of randomness should be included to produce re-calibrated scores that can reasonably be interpreted as the players’ achievements.     

Exploring teachers' instructional design practices from a systems design perspective

Curricular changes in higher vocationaleducation have rendered teachers' instructionaldesign activities increasingly important. Usinga repertory grid technique, this paper sets outto analyse current design activities of tenteacher trainers. Their actual approach iscompared with an instructional systems design(ISD) approach and related to innovativeteacher roles. Teachers' activities show animbalance in two ID phases, that is problemanalysis and evaluation. The results suggestthat they attempt to translate curricular goalsdirectly into concrete lessons and they payrelatively little attention to evaluation. Inline with this finding, they underrate the twoinnovative teacher roles of the `diagnostician'and the `evaluator'. It is argued thatimbalanced or incomplete design approaches andperceived roles may hinder innovation ineducation. Implications for the support ofteachers' design activities are discussed.