No abstract

No abstract

Silicone-based polymers have been widely used for many applications, but their extremely low surface energies and the resulting poor adhesion have been the cause for continuous problems. Herein, a novel adhesion improvement technique using an interlocked finger structure is demonstrated, which enables up to 24.8 and 7.3-fold increases in adhesion compared to the untreated and conventional plasma-treated cases, respectively. The interlocked finger structure is fabricated by surface-confined dissolution and subsequent directional melt crystallization of a solvent. After removing the solvent crystals, porous surfaces are prepared from polyurethane, polyvinyl alcohol, and polystyrene, and these are used to fabricate interfaces of interlocked finger structures with polydimethylsiloxane. The improvement in adhesion strength linearly depends on the pore depth of the prepared surfaces. This novel technique of surface adhesion could improve the performance of polymers with intrinsically poor adhesion in future applications.

Recently, crystallization engineering has become a novel processing technique for various materials, including ceramics, polymers, and composites. Herein, a novel processing technology of polymers based on controlled directional crystallization was developed for biomimetic surfaces and interfaces. Solvent was allowed to come into contact with a polymer surface for a limited time, followed by controlled crystallization of the solvent along a temperature gradient perpendicular to the surface. As a result, perpendicular pores of well-defined patterns were successfully prepared, as well as adhesive-free strong interfaces mimicking neuromuscular junctions. By increasing the temperature of the polymer or solvent contact time, the pore depth and contact angle increased. Highly hydrophobic surfaces of polycarbonate were efficiently prepared, and interfacial adhesion with polydimethylsiloxane was improved by more than 4-fold. This novel processing technique based on crystal engineering could open completely new application possibilities, particularly for biomedical devices, soft lithography, microfabrication, soft sensors, and flexible and stretchable electronics.

No abstract

No abstract

Student assessment is one of the most fundamental tasks in the field of AI Education (AIEd). One of the most common approach to student assessment is Knowledge Tracing (KT), which evaluates a student's knowledge state by predicting whether the student will answer a given question correctly or not. However, in the context of multiple choice (polytomous) questions, conventional KT approaches are limited in that they only consider the binary (dichotomous) correctness label (i.e., correct or incorrect), and disregard the specific option chosen by the student. Meanwhile, Option Tracing (OT) attempts to model a student by predicting which option they will choose for a given question, but overlooks the correctness information. In this paper, we propose Dichotomous-Polytomous Multi-Task Learning (DP-MTL), a multi-task learning framework that combines KT and OT for more precise student assessment. In particular, we show that the KT objective acts as a regularization term for OT in the DP-MTL framework, and propose an appropriate architecture for applying our method on top of existing deep learning-based KT models. We experimentally confirm that DP-MTL significantly improves both KT and OT performances, and also benefits downstream tasks such as Score Prediction (SP).

Student assessment is one of the most fundamental tasks in the field of AI Education (AIEd). One of the most common approach to student assessment is Knowledge Tracing (KT), which evaluates a student's knowledge state by predicting whether the student will answer a given question correctly or not. However, in the context of multiple choice (polytomous) questions, conventional KT approaches are limited in that they only consider the binary (dichotomous) correctness label (i.e., correct or incorrect), and disregard the specific option chosen by the student. Meanwhile, Option Tracing (OT) attempts to model a student by predicting which option they will choose for a given question, but overlooks the correctness information. In this paper, we propose Dichotomous-Polytomous Multi-Task Learning (DP-MTL), a multi-task learning framework that combines KT and OT for more precise student assessment. In particular, we show that the KT objective acts as a regularization term for OT in the DP-MTL framework, and propose an appropriate architecture for applying our method on top of existing deep learning-based KT models. We experimentally confirm that DP-MTL significantly improves both KT and OT performances, and also benefits downstream tasks such as Score Prediction (SP).