Employing large-scale pre-trained model CLIP to conduct video-text retrieval task (VTR) has become a new trend, which exceeds previous VTR methods. Though, due to the heterogeneity of structures and contents between video and text, previous CLIP-based models are prone to overfitting in the training phase, resulting in relatively poor retrieval performance. In this paper, we propose a multi-stream Corpus Alignment network with single gate Mixture-of-Experts (CAMoE) and a novel Dual Softmax Loss (DSL) to solve the two heterogeneity. The CAMoE employs Mixture-of-Experts (MoE) to extract multi-perspective video representations, including action, entity, scene, etc., then align them with the corresponding part of the text. In this stage, we conduct massive explorations towards the feature extraction module and feature alignment module, and conclude an efficient VTR framework. DSL is proposed to avoid the oneway optimum-match which occurs in previous contrastive methods. Introducing the intrinsic prior of each pair in a batch, DSL serves as a reviser to correct the similarity matrix and achieves the dual optimal match. DSL is easy to implement with only one-line code but improves significantly. The results show that the proposed CAMoE and DSL are of strong efficiency, and each of them is capable of achieving State-of-The-Art (SOTA) individually on various benchmarks such as MSR-VTT, MSVD, and LSMDC. Further, with both of them, the performance is advanced to a great extent, surpassing the previous SOTA methods for around 4.6% R@1 in MSR-VTT. The code will be available soon at https://github.com/starmemda/CAMoE/
The key steps of a transfer of two-dimensional (2D) materials are the delamination of the as-grown material from a growth substrate and the lamination of the 2D material on a target substrate. In state-of-the-art transfer experiments, these steps remain very challenging, and transfer variations often result in unreliable 2D material properties. Here, it is demonstrated that interfacial water can insert between graphene and its growth substrate despite the hydrophobic behavior of graphene. It is understood that interfacial water is essential for an electrochemistry-based graphene delamination from a Pt surface. Additionally, the lamination of graphene to a target wafer is hindered by intercalation effects, which can even result in graphene delamination from the target wafer. For circumvention of these issues, a direct, support-free graphene transfer process is demonstrated, which relies on the formation of interfacial water between graphene and its growth surface, while avoiding water intercalation between graphene and the target wafer by using hydrophobic silane layers on the target wafer. The proposed direct graphene transfer also avoids polymer contamination (no temporary support layer) and eliminates the need for etching of the catalyst metal. Therefore, recycling of the growth template becomes feasible. The proposed transfer process might even open the door for the suggested atomic-scale interlocking-toy-brick-based stacking of different 2D materials, which will enable a more reliable fabrication of van der Waals heterostructure-based devices and applications.
In view of its epitaxial seeding capability, c-plane single crystalline sapphire represents one of the most enticing, industry-compatible templates to realize manufacturable deposition of single crystalline two-dimensional transition metal dichalcogenides (MX 2 ) for functional, ultrascaled, nanoelectronic devices beyond silicon. Despite sapphire being atomically flat, the surface topography, structure, and chemical termination vary between sapphire terraces during the fabrication process. To date, it remains poorly understood how these sapphire surface anomalies affect the local epitaxial registry and the intrinsic electrical properties of the deposited MX 2 monolayer. Therefore, molybdenum disulfide (MoS 2 ) is deposited by metal−organic chemical vapor deposition (MOCVD) in an industry-standard epitaxial reactor on two types of c-plane sapphire with distinctly different terrace and step dimensions. Complementary scanning probe microscopy techniques reveal an inhomogeneous conductivity profile in the first epitaxial MoS 2 monolayer on both sapphire templates. MoS 2 regions with poor conductivity correspond to sapphire terraces with uncontrolled topography and surface structure. By intentionally applying a substantial off-axis cut angle (1°in this work), the sapphire terrace width and step heightand thus also surface structurebecome more uniform across the substrate and MoS 2 conducts the current more homogeneously. Moreover, these effects propagate into the extrinsic MoS 2 device performance: the field-effect transistor variability reduces both within and across wafers at higher median electron mobility. Carefully controlling the sapphire surface topography and structure proves an essential prerequisite to systematically study and control the MX 2 growth behavior and capture the influence on its structural and electrical properties.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.