Product review summarization through question retrieval and diversification

Product reviews have become an important resource for customers before they make purchase decisions. However, the abundance of reviews makes it difficult for customers to digest them and make informed choices. In our study, we aim to help customers who want to quickly capture the main idea of a lengthy product review before they read the details. In contrast with existing work on review analysis and document summarization, we aim to retrieve a set of real-world user questions to summarize a review. In this way, users would know what questions a given review can address and they may further read the review only if they have similar questions about the product. Specifically, we design a two-stage approach which consists of question selection and question diversification. For question selection phase, we first employ probabilistic retrieval models to locate candidate questions that are relevant to a given review. A Recurrent Neural Network Encoder–Decoder is utilized to measure the “answerability” of questions to a review. We then design a set function to re-rank the questions with the goal of rewarding diversity in the final question set. The set function satisfies submodularity and monotonicity, which results in an efficient greedy algorithm of submodular optimization. Evaluation on product reviews from two categories shows that the proposed approach is effective for discovering meaningful questions that are representative of individual reviews.     

Learning mathematics from worked-out examples: Analyzing and fostering self-explanations

Recent research has shown that learning from worked-out examples is of major importance for initial skill acquisition in well-structured domains such as mathematics. However, only those learners who actively process the presented examples profit noticeably from this learning mode. Specifically, the learning outcomes depend on how well the learners explain the solution steps presented in the examples to themselves (‘self-explanation effect”). In a series of studies on learning mathematics from examples, learners’ spontaneous self-explanations and instructional means used to encourage self-explanations were investigated. In this research, the following main findings were obtained. Most learners were rather passive with respect to their spontaneous self-explanations. Among the active and successful learners, two subgroups employing different self-explanation styles could be identified. With regard to the instructional means used to induce effective example processing, it turned out that to employ “learning by teaching” in order to stimulate explanation activities was of very limited use. Attempts to directly train for or elicit certain types of self-explanations were more successful. However, even in the latter case, self-explanations had inherent deficits (e.g., proneness to errors). Thus, we sought to design learning arrangements that try to integrate self-explanations with well-timed and well-adapted instructional explanations (e.g., from tutors) in order to enhance students’ problem-solving skills.     

Vertical microbubble column–A photonic lab-on-chip for cultivation and online analysis of yeast cell cultures

This paper presents a vertically positioned microfluidic system made of poly(dimethylsiloxane) (PDMS) and glass, which can be applied as a microbubble column (μBC) for biotechnological screening in suspension. In this μBC, microbubbles are produced in a cultivation chamber through an integrated nozzle structure. Thus, homogeneous suspension of biomass is achieved in the cultivation chamber without requiring additional mixing elements. Moreover, blockage due to produced carbon dioxide by the microorganisms—a problem predominant in common, horizontally positioned microbioreactors (MBRs)—is avoided, as the gas bubbles are released by buoyancy at the upper part of the microsystem. The patterned PDMS layer is based on an optimized two-lithographic process. Since the naturally hydrophobic PDMS causes problems for the sufficient production of microbubbles, a method based on polyelectrolyte multilayers is applied in order to allow continuous hydrophilization of the already bonded PDMS-glass-system. The μBC comprises various microelements, including stabilization of temperature, control of continuous bubble formation, and two optical configurations for measurement of optical density with two different sensitivities. In addition, the simple and robust application and handling of the μBC is achieved via a custom-made modular plug-in adapter. To validate the scalability from laboratory scale to microscale, and thus to demonstrate the successful application of the μBC as a screening instrument, a batch cultivation of Saccharomyces cerevisiae is performed in the μBC and compared to shake flask cultivation. Monitoring of the biomass growth in the μBC with the integrated online analytics resulted in a specific growth rate of 0.32 h⁻¹, which is almost identical to the one achieved in the shake flask cultivation (0.31 h⁻¹). Therefore, the validity of the μBC as an alternative screening tool compared to other conventional laboratory scale systems in bioprocess development is proven. In addition, vertically positioned microbioreactors show high potential in comparison to conventional screening tools, since they allow for high density of integrated online analytics and therefore minimize time and cost for screening and guarantee improved control and analysis of cultivation parameters.     

The Benefits of Peer Collaboration: A Replication with a Delayed Posttest

The benefits of peer collaboration on strategy use, metacognitive causal attribution, and recall were supported in an earlier study (Manion & Alexander, 1997). Following that study two primary questions remained: Would the observed benefits last beyond an immediate posttest, and would they last regardless of which strategy the children chose to use? The current investigation explores those questions. Thirty-one fourth-grade students from two public schools in Sydney, Australia, experienced either a peer-collaborative treatment condition or an individual task (control condition). A posttest was conducted 5.5 weeks later. As was the case in the original study, children in the treatment condition outperformed children in the control condition in strategy use, metacognitive understanding of strategy chosen, and recall gain. Benefits also sustained regardless of the strategy children chose to implement on the task. Most importantly, the benefits lasted after a period of over a month. Implications for future research are discussed.     

Benefits of extracurricular participation in dissection in a prosection‐based medical anatomy program

The purpose of this study was to evaluate the extracurricular cadaveric dissection program available to medical students at an institution with a modern (time‐compressed, student‐centered, and prosection‐based) approach to medical anatomy education. Quantitative (Likert‐style questions) and qualitative data (thematic analysis of open‐ended commentary) were collated from a survey of three medical student cohorts who had completed preclerkship. Perceived benefits of dissection included the hands‐on learning style and the development of anatomy expertise, while the main barrier that limited participation was the time‐intensive nature of dissection. Despite perceived benefits, students preferred that dissection remain optional. Analysis of assessments for the MD2016 cohort revealed that dissection participation was associated with enhanced performance on anatomy items in each systems‐based unit examination, with the largest benefits observed on discriminating items that assessed knowledge application. In conclusion, this study revealed that there are academic and perceived benefits of extracurricular participation in dissection. While millennial medical students recognized these benefits, these students also indicated strong preference for having flexibility and choice in their anatomy education, including the choice to participate in cadaveric dissection. Anat Sci Educ 11: 294–302. © 2017 American Association of Anatomists.     

Verteilte Metadatenverwaltung und Anfragebearbeitung für Internet-Datenquellen

Zusammenfassung.   Wir pr?sentieren in dieser Arbeit das ObjectGlobe-System, ein offenes und verteiltes Anfragebearbeitungssystem auf Internet-Datenquellen. ObjectGlobe erweitert die begrenzten F?higkeiten des Internets bezüglich Anfragebearbeitung, indem es einen offenen Marktplatz schafft, in dem eine Vielzahl von Anbietern Rechenzeit, Daten und Anfrageoperatoren bereitstellen. Den Schwerpunkt dieses Beitrags bildet die im Rahmen des ObjectGlobe-Projekts entwickelte Metadatenverwaltung MDV, ein verteiltes Metadaten-Management-System. Es besitzt eine 3-schichtige Architektur und unterstützt Caching und Replikation in der Mittelschicht, so dass Anfragen lokal ausgewertet werden k?nnen. Benutzer spezifizieren die Informationen, die sie ben?tigen und die repliziert werden, mit Hilfe einer besonderen Regelsprache. Um Replikate aktuell zu halten und die Replikation von neuen und relevanten Informationen anzusto?en, verwendet das MDV-System einen neuartigen Publish & Subscribe-Algorithmus. Wir beschreiben diesen Algorithmus im Detail und zeigen, wie er mit Hilfe eines normalen relationalen Datenbanksystems implementiert werden kann. Eingegangen am 19. November 2001 / Angenommen am 30. Juli 2002