471 research outputs found
Task Execution Latency Minimization-based Joint Computation Offloading and Cell Selection for MEC-Enabled HetNets
No full textMulti-access edge computing (MEC) is a promising platform for cellular heterogeneous networks (HetNets). It brings the computation services of the core cloud into the nearest point of mobile devices (MDs). This paper jointly studies the problem of computation offloading and cell selection strategies in MEC-enabled HetNets under the assumption that MD can select one of the three ways to execute its task, i.e., complete local computing mode, complete MEC computing mode or partial computing mode. The problem of joint computation offloading and cell selection is formulated as a task execution latency minimization problem. This formulated optimization situation is NP-Hard, which is inconvenient to solve. Hence, we deduce the problem into two subproblems which is computation offloading subproblem and cell selection subproblem and propose extensive search method and Khun-Munkres algorithm respectively to obtain the optimal solution. Numerical results show the efficacy of our proposed algorithm
IntelLabs/bayesian-torch: Bayesian-Torch 0.4.0
No full text<p>New feature: Quantization framework in Bayesian-Torch.
Support Post Training Quantization of Bayesian deep neural networks, enables efficient INT8 inference on Intel platforms.</p>
What's Changed
<ul>
<li>Add support for output padding in flipout layers by @junliang-lin in <a href="https://github.com/IntelLabs/bayesian-torch/pull/20">https://github.com/IntelLabs/bayesian-torch/pull/20</a></li>
<li>Add support for Bayesian neural network quantization by @junliang-lin in <a href="https://github.com/IntelLabs/bayesian-torch/pull/23">https://github.com/IntelLabs/bayesian-torch/pull/23</a></li>
</ul>
New Contributors
<ul>
<li>@junliang-lin made their first contribution in <a href="https://github.com/IntelLabs/bayesian-torch/pull/20">https://github.com/IntelLabs/bayesian-torch/pull/20</a></li>
</ul>
<p><strong>Full Changelog</strong>: <a href="https://github.com/IntelLabs/bayesian-torch/compare/v0.3.0...v0.4.0">https://github.com/IntelLabs/bayesian-torch/compare/v0.3.0...v0.4.0</a></p>
On the group of zero-cycles of holomorphic symplectic varieties
No full textFor a moduli space of Bridgeland-stable objects on a K3 surface, we show that
the Chow class of a point is determined by the Chern class of the corresponding
object on the surface. This establishes a conjecture of Junliang Shen, Qizheng
Yin, and the second author
In-situ conversion of Al-SBA-15 wall structures to monolithic ZSM-5 nanoparticle aggregates
No full textThe amorphous wall structure of Al-SBA-15 was converted in-situ to uniform ZMS-5 nanocrystal aggregates. Si/Al ratio in resulting ZSM-5 nanocrystal aggregates is close to that of raw Al-SBA-15. The ZSM-5 aggregates exhibits a monolithic morphology, a replica of the shape of raw Al-SBA-15, which significantly allows the avoidance of a high-speed centrifugal separation process. The incompletely decomposed structure-directing template of Al-SBA-15 works as inert units inserting in the mesopores of Al-SBA-15 materials. The nucleation and growth of zeolite were physically confined inside a nano-scale space surrounded by the chemically inert units that were generated from the controlled decomposition of the surfactant.National Natural Science Foundation of China [51202134, 51472256]; Shandong Province Young and Middle-Aged Scientists Research Awards Fund [BS2012CL027]SCI(E)ARTICLE382-38518
A Transfer Report on the Development of a Framework to Evaluate Search Interfaces for their Support of Different User Types and Search Tactics
No full textAs the understanding of search systems, user needs and seeking strategies is developing, the design of search user interfaces is evolving to support more complicated and exploratory forms of search. With the design of new search features that enable these richer modes of exploration, comes the need to better understand the support they provide. In this report a new evaluation framework is presented that analyses search features for how they a) contribute to an overall interface, b) allow users to carry out different search tactics, and c) support different types of users and their needs. The novel contributions of the framework improve on some of the limitations of typical user studies, and allow search systems to be systematically analysed in much more detail and in much less time. The presented evaluation framework is then validated in three ways. First the validity of the models used as the building blocks of the framework are investigated through related work. Second the method of integrating these building-block models is validated and strengthened by consensus of expert opinion. Third, the overall approach is validated by comparing its analyses to the results of previously carried out user studies. The validation process has shown both the value of the framework and identified areas of future work that should be addressed for the framework to be completed. This report concludes with the set of contributions that the framework makes, and why the remaining work will be challenging, but critical to the final design
Souls of the Wind
No full textsynopsisAs the Guardian of the Four Winds, use your unique wind powers to complete the trials of the Wind Temples and free the Souls of the Wind from corruption. This is a tale of a world full of wind magic, through a beautiful 3rd person puzzle-adventure game crafted by Wandering Spirit Games.; ; Team Members:; Aaron Kapral - Character Artist and Animator; Azhar Siddiqui - Gameplay Programmer; Bradley Dawn - Level & Game Designer; Casey Gatlin - Lead Engineer and Creative Director; Conner Workman - Level Designer; Deeprashmi Nath - Producer and Game Designer; Dong Wang - Gameplay Programmer; Jennifer Egan - Narrative and Level Designer; Jolie Uk - VFX and Shader Artist; Junliang Zhang - Lighting and Technical Artist; Kevin Le - Gameplay Programmer; Manas Rawat - Gameplay and Tools Programmer; Mingxi Zou - UX/UI Artist; Nero Cao - Level and Game Designer; Saaransh Vatsa - Gameplay Programmer; Darianne Salinas - Social Media Specialist; Anita Lin - Audio DesignerPC; WindowsAdventure; Free to Play; RP
Magnetically driven formation of 3D freestanding soft bioscaffolds.
No full text3D soft bioscaffolds have great promise in tissue engineering, biohybrid robotics, and organ-on-a-chip engineering applications. Though emerging three-dimensional (3D) printing techniques offer versatility for assembling soft biomaterials, challenges persist in overcoming the deformation or collapse of delicate 3D structures during fabrication, especially for overhanging or thin features. This study introduces a magnet-assisted fabrication strategy that uses a magnetic field to trigger shape morphing and provide remote temporary support, enabling the straightforward creation of soft bioscaffolds with overhangs and thin-walled structures in 3D. We demonstrate the versatility and effectiveness of our strategy through the fabrication of bioscaffolds that replicate the complex 3D topology of branching vascular systems. Furthermore, we engineered hydrogel-based bioscaffolds to support biohybrid soft actuators capable of walking motion triggered by cardiomyocytes. This approach opens new possibilities for shaping hydrogel materials into complex 3D morphologies, which will further empower a broad range of biomedical applications
Puffball-inspired microrobotic systems with robust payload, strong protection, and targeted locomotion for on-demand drug delivery
No full textMicrorobots have been recognized as transformative solutions for drug delivery systems (DDSs) because they can navigate through the body to specific locations and enable targeted drug release. However, their realization is substantially limited by insufficient payload capacity, unavoidable drug leakage/deactivation, and strict modification/stability criteria for drugs. Natural puffballs possess fascinating features that are highly desirable for DDSs, including a large fruitbody for storing spores, a flexible protective cap, and environmentally-triggered release mechanisms. This report presents a puffball-inspired microrobotic system which incorporates: an internal chamber for loading large drug quantities and spatial drug separation; and a near-infrared-responsive top-sealing layer offering strong drug protection and on-demand release. These puffball-inspired microrobots (PIMs) display tunable loading capacities up to high concentrations and enhanced drug protection with minimal drug leakage.
Upon near-infrared laser irradiation, on-demand drug delivery with rapid release efficiency is achieved. The PIMs also demonstrate translational motion velocities, switchable motion modes, and precise locomotion under a rotating magnetic field. This work provides strong proof-of-concept for a DDS that combines the superior locomotion capability of microrobots with the
unique characteristics of puffballs, thereby illustrating a versatile avenue for development of a new generation of microrobots for targeted drug delivery
Synchronization Motion Control for Electro-hydraulic System of Tensioner for Pipelaying Vessel
No full text- …
