Hashtag-based topic evolution in social media

The rise of online social media has led to an explosion of metadata-containing user generated content. The tracking of metadata distribution is essential to understand social media. This paper presents two statistical models that detect interpretable topics over time along with their hashtags distribution. A topic is represented by a cluster of words that frequently occur together, and a context is represented by a cluster of hashtags, i.e., the hashtag distribution. The models combine a context with a related topic by jointly modeling words with hashtags and time. Experiments with real-world datasets demonstrate that the proposed models discover topics over time with related contexts effectively.     

A transfer cost-sensitive boosting approach for cross-project defect prediction

Software defect prediction has been regarded as one of the crucial tasks to improve software quality by effectively allocating valuable resources to fault-prone modules. It is necessary to have a sufficient set of historical data for building a predictor. Without a set of sufficient historical data within a company, cross-project defect prediction (CPDP) can be employed where data from other companies are used to build predictors. In such cases, a transfer learning technique, which extracts common knowledge from source projects and transfers it to a target project, can be used to enhance the prediction performance. There exists the class imbalance problem, which causes difficulties for the learner to predict defects. The main impacts of imbalanced data under cross-project settings have not been investigated in depth. We propose a transfer cost-sensitive boosting method that considers both knowledge transfer and class imbalance for CPDP when given a small amount of labeled target data. The proposed approach performs boosting that assigns weights to the training instances with consideration of both distributional characteristics and the class imbalance. Through comparative experiments with the transfer learning and the class imbalance learning techniques, we show that the proposed model provides significantly higher defect detection accuracy while retaining better overall performance. As a result, a combination of transfer learning and class imbalance learning is highly effective for improving the prediction performance under cross-project settings. The proposed approach will help to design an effective prediction model for CPDP. The improved defect prediction performance could help to direct software quality assurance activities and reduce costs. Consequently, the quality of software can be managed effectively.     

Collaborative object-oriented visualization environment

In this paper, we present a Collaborative Object-oriented Visualization Environment (COVE) which provides a flexible and extensible framework for collaborative visualization. COVE integrates collaborative and parallel computing environments based on a distributed object model. It is built as a collection of concurrent objects: collaborative and application objects which interact with one another to construct collaborative parallel computing environments. The former enables COVE to execute various collaborative functions, while the latter allows it to execute fast parallel visualization in various modes. Also, flexibility and extensibility are provided by plugging the proper application objects into COVE at run-time, and making them interact with one another through collaboration objects. For our experiment, three visualization modes for volume rendering are designed and implemented to support the fast and flexible analysis of volume data in a collaborative environment. This work has been supported by KIPA-Information Technology Research Center, University research program by Ministry of Information & Communication, and Brain Korea 21 projects in 2005.     

Advanced photospacers with elastic recovery

A new series of photospacers has been prepared with different types of crosslinkers to improve elastic recovery. Afunctional crosslinker with six reactive groups demonstrates the best elastic recovery. As the quantity of crosslinker is increased, the elastic recovery also increases, probably due to an increase in the crosslinking density. The use of ADMS EPS® results in high resolution, good uniformity, and high production yield in the liquid‐crystal‐display (LCD) process. Especially, EPS® improves and solves problems such as viewing angle, crosstalk, and dark spots.     

Queue-based local scheduling and global coordination for real-time operation control in a container terminal

To maximize the productivity of a container terminal, the operations of various types of equipments should be optimized and synchronized in real time. However, use of optimization techniques such as mathematical programming or search-based meta-heuristics becomes difficult when given a large-scaled problem due to their high computational cost. Addressing this problem, the queue-based local scheduling and global coordination method proposed in this paper stands as a viable alternative. The method consists of the following steps. First, separate schedules are locally generated for each equipment type using a queue-based dispatching heuristic which pays attention to the queue lengths of the quay cranes (QCs) under service. Next, the schedules are executed via a simulation and a notable QC delay is identified. Based on the analysis on the causes of this delay, some compromising adjustments are made to the priorities of relevant jobs. Then, the localized scheduling followed by the adjustment is repeated until the termination condition is met. Adopting simple heuristics in the local scheduling phase, the overall process easily meets the real-time constraint, yet producing an integrated schedule with a better global perspective than the myopic heuristic-only approach.     

Connolly Surface on an Atomic Structure via Voronoi Diagram of Atoms

One of the most important geometric structures of a protein is the Connolly surface of protein since a Connolly surface plays an important role in protein folding, docking, interactions between proteins, amongst other things. This paper presents an algorithm for precisely and efficiently computing the Connolly surface of a protein using a proposed geometric construct called β-shape based on the Voronoi diagram of atoms in the protein. Given the Voronoi diagram of atoms based on the Euclidean distance from the atom surfaces, the proposed algorithm first computes a β-shape with an appropriate probe. Then, the Connolly surface is computed by employing the blending operation on the atomic complex of the protein by the given probe.     

Reliable Piezoelectricity in Bilayer WSe2 for Piezoelectric Nanogenerators

Recently, piezoelectricity has been observed in 2D atomically thin materials, such as hexagonal‐boron nitride, graphene, and transition metal dichalcogenides (TMDs). Specifically, exfoliated monolayer MoS₂ exhibits a high piezoelectricity that is comparable to that of traditional piezoelectric materials. However, monolayer TMD materials are not regarded as suitable for actual piezoelectric devices due to their insufficient mechanical durability for sustained operation while Bernal‐stacked bilayer TMD materials lose noncentrosymmetry and consequently piezoelectricity. Here, it is shown that WSe₂ bilayers fabricated via turbostratic stacking have reliable piezoelectric properties that cannot be obtained from a mechanically exfoliated WSe₂ bilayer with Bernal stacking. Turbostratic stacking refers to the transfer of each chemical vapor deposition (CVD)‐grown WSe₂ monolayer to allow for an increase in degrees of freedom in the bilayer symmetry, leading to noncentrosymmetry in the bilayers. In contrast, CVD‐grown WSe₂ bilayers exhibit very weak piezoelectricity because of the energetics and crystallographic orientation. The flexible piezoelectric WSe₂ bilayers exhibit a prominent mechanical durability of up to 0.95% of strain as well as reliable energy harvesting performance, which is adequate to drive a small liquid crystal display without external energy sources, in contrast to monolayer WSe₂ for which the device performance becomes degraded above a strain of 0.63%.     

Practical considerations of Si-based anodes for lithium-ion battery applications

Using Si-based anodes in Li-ion batteries is one of the most feasible approaches to achieve high energy densities despite their disadvantages, such as low conductivity and massive volume expansion, which cause unstable solid electrolyte interphase layers with mechanical failure. The forefront in research and development to address the above challenges suggests the possibility of fully commercially viable cells using various structural and interfacial modifications. In particular, we present a discussion of each dimension of Si-based anodes in multiple controlled systems, including plain, hollow, porous, and uniquely engineered structures, which are further evaluated based on their anode performances, such as initial reversibility, capacity retention for extended cycles with its efficiency, degree of volume expansion tolerance, and rate capabilities, by several practical standards in half cells. With these practical considerations, multi-dimensional structures with uniform size distributions (micrometers, on average) are strongly desired to satisfy the rigorous requirements for widespread applications. Furthermore, we closely examined several full cells composed of Si-based multicomponent anodes coupled with suitable cathodes based on practical standards to propose future research directions for Si-based anodes to keep pace with the rapidly changing market demands for diverse energy storage systems.

Open image in new window