-
Power Efficient Video Super-Resolution on Mobile NPUs with Deep Learning, Mobile AI & AIM 2022 challenge: Report
Authors:
Andrey Ignatov,
Radu Timofte,
Cheng-Ming Chiang,
Hsien-Kai Kuo,
Yu-Syuan Xu,
Man-Yu Lee,
Allen Lu,
Chia-Ming Cheng,
Chih-Cheng Chen,
Jia-Ying Yong,
Hong-Han Shuai,
Wen-Huang Cheng,
Zhuang Jia,
Tianyu Xu,
Yijian Zhang,
Long Bao,
Heng Sun,
Diankai Zhang,
Si Gao,
Shaoli Liu,
Biao Wu,
Xiaofeng Zhang,
Chengjian Zheng,
Kaidi Lu,
Ning Wang
, et al. (29 additional authors not shown)
Abstract:
Video super-resolution is one of the most popular tasks on mobile devices, being widely used for an automatic improvement of low-bitrate and low-resolution video streams. While numerous solutions have been proposed for this problem, they are usually quite computationally demanding, demonstrating low FPS rates and power efficiency on mobile devices. In this Mobile AI challenge, we address this prob…
▽ More
Video super-resolution is one of the most popular tasks on mobile devices, being widely used for an automatic improvement of low-bitrate and low-resolution video streams. While numerous solutions have been proposed for this problem, they are usually quite computationally demanding, demonstrating low FPS rates and power efficiency on mobile devices. In this Mobile AI challenge, we address this problem and propose the participants to design an end-to-end real-time video super-resolution solution for mobile NPUs optimized for low energy consumption. The participants were provided with the REDS training dataset containing video sequences for a 4X video upscaling task. The runtime and power efficiency of all models was evaluated on the powerful MediaTek Dimensity 9000 platform with a dedicated AI processing unit capable of accelerating floating-point and quantized neural networks. All proposed solutions are fully compatible with the above NPU, demonstrating an up to 500 FPS rate and 0.2 [Watt / 30 FPS] power consumption. A detailed description of all models developed in the challenge is provided in this paper.
△ Less
Submitted 7 November, 2022;
originally announced November 2022.
-
On-the-Fly, Sample-Tailored Optimisation of NMR Experiments
Authors:
Jonathan R. J. Yong,
Mohammadali Foroozandeh
Abstract:
NMR experiments, indispensable to chemists in many areas of research, are often run with generic, unoptimised experimental parameters. This approach makes robust and automated acquisition on different samples and instruments extremely challenging. Here, we introduce NMR-POISE (Parameter Optimisation by Iterative Spectral Evaluation), the first demonstration of on-the-fly, sample-tailored, and full…
▽ More
NMR experiments, indispensable to chemists in many areas of research, are often run with generic, unoptimised experimental parameters. This approach makes robust and automated acquisition on different samples and instruments extremely challenging. Here, we introduce NMR-POISE (Parameter Optimisation by Iterative Spectral Evaluation), the first demonstration of on-the-fly, sample-tailored, and fully automated optimisation of a wide range of NMR experiments. We illustrate how POISE maximises spectral sensitivity and quality with a diverse set of 1D and 2D examples, ranging from HSQC and NOESY experiments to ultrafast and pure shift techniques. Our Python implementation of POISE has an interface integrated into Bruker's TopSpin software, one of the most widely used platforms for NMR acquisition and automation, allowing NMR optimisations to be run without direct user supervision. We predict that POISE will find widespread usage in academia and industry, where sample-specific and automated experiment optimisation is mandatory.
△ Less
Submitted 16 August, 2021;
originally announced August 2021.
-
Anti-lock Brake System for Integrated Electric Parking Brake Actuator Based on Sliding-mode Control
Authors:
Dongliang Wang,
Yiyong Yang,
Wei Yu,
Jiawang Yong,
Xiaoxu Dong
Abstract:
Integrated electric parking brake (iEPB) is popularizing on passenger cars due to its easier operation and automatic functions. As a parking brake, EPB have to act as the secondary brake system in case of hydraulic brake failure. To guarantee the stability and safety of a car during iEPB braking, the rear slip ratio has to be controlled accurately within the optimized value to get the shortest bra…
▽ More
Integrated electric parking brake (iEPB) is popularizing on passenger cars due to its easier operation and automatic functions. As a parking brake, EPB have to act as the secondary brake system in case of hydraulic brake failure. To guarantee the stability and safety of a car during iEPB braking, the rear slip ratio has to be controlled accurately within the optimized value to get the shortest brake distance without undesired loss of control. In this paper, a sliding-mode controller (SMC) is investigated to achieve rear-wheel anti-lock brake control, which is robust against uncertainties and disturbance of the parameters. And a sliding-mode observer (SMO) is present to estimate the load torque of d.c. motor and calculate the brake torque. The tyre/road friction coefficient estimator is designed to obtain the optimal rear slip ratio timely. The simulation model of iEPB system is initially constructed in AMESim and the vehicle model is built in MATLAB/Simulink, and the complete system is co-simulated by these two software simultaneously with different road conditions. Simulation results show that the proposed observer and estimator are feasible. This study may provide a useful method to realize rear slip ratio control so that the safety and stability of vehicle could be improved significantly in specified condition
△ Less
Submitted 8 November, 2018; v1 submitted 25 October, 2018;
originally announced October 2018.
-
Hamilton-Jacobi Equations and Two-Person Zero-Sum Differential Games with Unbounded Controls
Authors:
Hong Qiu,
Jiongmin Yong
Abstract:
A two-person zero-sum differential game with unbounded controls is considered. Under proper coercivity conditions, the upper and lower value functions are characterized as the unique viscosity solutions to the corresponding upper and lower Hamilton--Jacobi--Isaacs equations, respectively. Consequently, when the Isaacs' condition is satisfied, the upper and lower value functions coincide, leading t…
▽ More
A two-person zero-sum differential game with unbounded controls is considered. Under proper coercivity conditions, the upper and lower value functions are characterized as the unique viscosity solutions to the corresponding upper and lower Hamilton--Jacobi--Isaacs equations, respectively. Consequently, when the Isaacs' condition is satisfied, the upper and lower value functions coincide, leading to the existence of the value function. Due to the unboundedness of the controls, the corresponding upper and lower Hamiltonians grow super linearly in the gradient of the upper and lower value functions, respectively. A uniqueness theorem of viscosity solution to Hamilton--Jacobi equations involving such kind of Hamiltonian is proved, without relying on the convexity/concavity of the Hamiltonian. Also, it is shown that the assumed coercivity conditions guaranteeing the finiteness of the upper and lower value functions are sharp in some sense.
△ Less
Submitted 17 February, 2012; v1 submitted 5 July, 2011;
originally announced July 2011.
-
Mean-Field Backward Stochastic Volterra Integral Equations
Authors:
Yufeng Shi,
Tianxiao Wang,
Jiongmin Yong
Abstract:
Mean-field backward stochastic Volterra integral equations (MF-BSVIEs, for short) are introduced and studied. Well-posedness of MF-BSVIEs in the sense of introduced adapted M-solutions is established. Two duality principles between linear mean-field (forward) stochastic Volterra integral equations (MF-FSVIEs, for short) and MF-BSVIEs are obtained. As applications, a multi-dimensional comparison th…
▽ More
Mean-field backward stochastic Volterra integral equations (MF-BSVIEs, for short) are introduced and studied. Well-posedness of MF-BSVIEs in the sense of introduced adapted M-solutions is established. Two duality principles between linear mean-field (forward) stochastic Volterra integral equations (MF-FSVIEs, for short) and MF-BSVIEs are obtained. As applications, a multi-dimensional comparison theorem is proved for adapted M-solutions of MF-BSVIEs and a maximum principle is established for an optimal control of MF-FSVIEs.
△ Less
Submitted 5 July, 2011; v1 submitted 25 April, 2011;
originally announced April 2011.
