A Data-Driven Optimized Mechanism for Improving Online Collaborative Learning: Taking Cognitive Load into Account.

Research on online collaborative learning has explored various methods of collaborative improvement. Recently, learning analytics have been increasingly adopted for ascertaining learners' states and promoting collaborative performance. However, little effort has been made to investigate the transformation of collaborative states or to consider cognitive load as an essential factor for collaborative intervention. By bridging collaborative cognitive load theory and system dynamics modeling methods, this paper revealed the transformation of online learners' collaborative states through data analysis, and then proposed an optimized mechanism to ameliorate online collaboration. A quasi-experiment was conducted with 91 college students to examine the potential of the optimized mechanism in collaborative state transformation, awareness of collaboration, learning achievement, and cognitive load. The promising results demonstrated that students learning with the optimized mechanism performed significantly differently in collaboration and knowledge acquisition, and no additional burden in cognitive load was noted.

Dual-Graph Attention Convolution Network for 3-D Point Cloud Classification.

Three-dimensional point cloud classification is fundamental but still challenging in 3-D vision. Existing graph-based deep learning methods fail to learn both low-level extrinsic and high-level intrinsic features together. These two levels of features are critical to improving classification accuracy. To this end, we propose a dual-graph attention convolution network (DGACN). The idea of DGACN is to use two types of graph attention convolution operations with a feedback graph feature fusion mechanism. Specifically, we exploit graph geometric attention convolution to capture low-level extrinsic features in 3-D space. Furthermore, we apply graph embedding attention convolution to learn multiscale low-level extrinsic and high-level intrinsic fused graph features together. Moreover, the points belonging to different parts in real-world 3-D point cloud objects are distinguished, which results in more robust performance for 3-D point cloud classification tasks than other competitive methods, in practice. Our extensive experimental results show that the proposed network achieves state-of-the-art performance on both the synthetic ModelNet40 and real-world ScanObjectNN datasets.

Identifying Learners' Interaction Patterns in an Online Learning Community.

The interactions among all members of an online learning community significantly impact collaborative reflection (co-reflection). Although the relationship between learners' roles and co-reflection levels has been explored by previous researchers, it remains unclear when and with whom learners at different co-reflection levels tend to interact. This study adopted multiple methods to examine the interaction patterns of diverse roles among learners with different co-reflection levels based on 11,912 posts. First, the deep learning technique was applied to assess learners' co-reflection levels. Then, a social network analysis (SNA) was conducted to identify the emergent roles of learners. Furthermore, a lag sequence analysis (LSA) was employed to reveal the interaction patterns of the emergent roles among learners with different co-reflection levels. The results showed that most learners in an online learning community reached an upper-middle co-reflection level while playing an inactive role in the co-reflection process. Moreover, higher-level learners were superior in dialog with various roles and were more involved in self-rethinking during the co-reflection process. In particular, they habitually began communication with peers and then with the teacher. Based on these findings, some implications for facilitating online co-reflection from the perspective of roles is also discussed.

Investigating the Influence of Interaction on Learning Persistence in Online Settings: Moderation or Mediation of Academic Emotions?

Learning persistence is a critical element for successful online learning. The evidence provided by psychologists and educators has shown that students' interaction (student-student (SS) interaction, student-instructor (SI) interaction, and student-content (SC) interaction) significantly affects their learning persistence, which is also related to their academic emotions. However, few studies explore the relations among students' interaction, academic emotions and learning persistence in online learning environments. Furthermore, no research has focused on multi-dimensional students' interaction and specific academic emotions. Based on person-environment interaction model and transactional distance theory, this study investigates the relationship between students' interaction and learning persistence from the perspective of moderation and mediation of academic emotions including enjoyment, boredom, and anxiety. Data were collected from 339 students who had online learning experience in China. AMOS 22.0 (IBM, Armonk, NY, USA) and SPSS 22.0 (IBM, Armonk, NY, USA) were employed to analyze the mediating and moderating effects of academic emotions, respectively. The results revealed that students' interaction and academic emotions directly related to learning persistence. Specifically, enjoyment, anxiety and boredom had significant mediating and moderating effects on the relationship between students' interaction and learning persistence. Based on these findings, we further discussed the theoretical and practical implications on how to facilitate students' learning persistence in online learning environments.

Metal oxide multilayer hard mask system for 3D nanofabrication.

We demonstrate the preparation and exploitation of multilayer metal oxide hard masks for lithography and 3D nanofabrication. Atomic layer deposition (ALD) and focused ion beam (FIB) technologies are applied for mask deposition and mask patterning, respectively. A combination of ALD and FIB was used and a patterning procedure was developed to avoid the ion beam defects commonly met when using FIB alone for microfabrication. ALD grown Al2O3/Ta2O5/Al2O3 thin film stacks were FIB milled with 30 keV gallium ions and chemically etched in 5% tetramethylammonium hydroxide at 50 °C. With metal evaporation, multilayers consisting of amorphous oxides Al2O3 and Ta2O5 can be tailored for use in 2D lift-off processing, in preparation of embedded sub-100 nm metal lines and for multilevel electrical contacts. Good pattern transfer was achieved by lift-off process from the 2D hard mask for micro- and nano-scaled fabrication. As a demonstration of the applicability of this method to 3D structures, self-supporting 3D Ta2O5 masks were made from a film stack on gold particles. Finally, thin film resistors were fabricated by utilizing controlled stiction of suspended Ta2O5 structures.

Selective etching of focused gallium ion beam implanted regions from silicon as a nanofabrication method.

A focused ion beam (FIB) is otherwise an efficient tool for nanofabrication of silicon structures but it suffers from the poor thermal stability of the milled surfaces caused by segregation of implanted gallium leading to severe surface roughening upon already slight annealing. In this paper we show that selective etching with KOH:H2O2 solutions removes the surface layer with high gallium concentration while blocking etching of the surrounding silicon and silicon below the implanted region. This remedies many of the issues associated with gallium FIB nanofabrication of silicon. After the gallium removal sub-nm surface roughness is retained even during annealing. As the etching step is self-limited to a depth of 25-30 nm for 30 keV ions, it is well suited for defining nanoscale features. In what is essentially a reversal of gallium resistless lithography, local implanted areas can be prepared and then subsequently etched away. Nanopore arrays and sub-100 nm trenches can be prepared this way. When protective oxide masks such as Al2O3 grown with atomic layer deposition are used together with FIB milling and KOH:H2O2 etching, ion-induced amorphization can be confined to sidewalls of milled trenches.

Combining focused ion beam and atomic layer deposition in nanostructure fabrication.

Combining the strengths of atomic layer deposition (ALD) with focused ion beam (FIB) milling provides new opportunities for making 3D nanostructures with flexible choice of materials. Such structures are of interest in prototyping microelectronic and MEMS devices which utilize ALD grown thin films. As-milled silicon structures suffer from segregation and roughening upon heating, however. ALD processes are typically performed at 200-500 °C, which makes thermal stability of the milled structures a critical issue. In this work Si substrates were milled with different gallium ion beam incident angles and then annealed at 250 °C. The amount of implanted gallium was found to rapidly decrease with increasing incident angle with respect of surface normal, which therefore improves the thermal stability of the milled features. 60° incident angle was found as the best compromise with respect to thermal stability and ease of milling. ALD Al2O3 growth at 250 °C on the gallium FIB milled silicon was possible in all cases, even when segregation was taking place. ALD Al2O3 could be used both for creating a chemically uniform surface and for controlled narrowing of FIB milled trenches.