Worst case vs. average performance estimates for unconstrained NMPC schemes

Shared cache interference in multi-core architectures has been recognized as one of major factors that degrade predictability of a mixed-critical real-time system. Due to the unpredictable cache interference, the behavior of shared cache is hard to predict and analyze statically in multi-core architectures executing mixed-critical tasks, which will not only result in difficulty of estimating the worst-case execution time (WCET) but also introduce significant worst-case timing penalties for critical tasks. Therefore, cache management in mixed-critical multi-core systems has become a challenging task. In this paper, we present a dynamic partitioned cache memory for mixed-critical real-time multi-core systems. In this architecture, critical tasks can dynamically allocate and release the cache resourse during the execution interval according to the real-time workload. This dynamic partitioned cache can, on the one hand, provide the predicable cache performance for critical tasks. On the other hand, the released cache can be dynamically used by non-critical tasks to improve their average performance. We demonstrate and prototype our system design on the embedded FPGA platform. Measurements from the prototype clearly demonstrate the benefits of the dynamic partitioned cache for mixed-critical real-time multi-core systems.

Download Full-text

Greedy Matching in Bipartite Random Graphs

Stochastic Systems ◽

10.1287/stsy.2021.0082 ◽

2021 ◽

Author(s):

Nick Arnosti

Keyword(s):

Random Graphs ◽

Bipartite Graphs ◽

Approximation Ratio ◽

Prior Work ◽

Asymptotic Results ◽

Worst Case ◽

Average Performance ◽

Finite Graphs ◽

Random Priority ◽

Priority Order

This paper studies the performance of greedy matching algorithms on bipartite graphs [Formula: see text]. We focus primarily on three classical algorithms: [Formula: see text], which sequentially selects random edges from [Formula: see text]; [Formula: see text], which sequentially matches random vertices in [Formula: see text] to random neighbors; and [Formula: see text], which generates a random priority order over vertices in [Formula: see text] and then sequentially matches random vertices in [Formula: see text] to their highest-priority remaining neighbor. Prior work has focused on identifying the worst-case approximation ratio for each algorithm. This guarantee is highest for [Formula: see text] and lowest for [Formula: see text]. Our work instead studies the average performance of these algorithms when the edge set [Formula: see text] is random. Our first result compares [Formula: see text] and [Formula: see text] and shows that on average, [Formula: see text] produces more matches. This result holds for finite graphs (in contrast to previous asymptotic results) and also applies to “many to one” matching in which each vertex in [Formula: see text] can match with multiple vertices in [Formula: see text]. Our second result compares [Formula: see text] and [Formula: see text] and shows that the better worst-case guarantee of [Formula: see text] does not translate into better average performance. In “one to one” settings where each vertex in [Formula: see text] can match with only one vertex in [Formula: see text], the algorithms result in the same number of matches. When each vertex in [Formula: see text] can match with two vertices in [Formula: see text] produces more matches than [Formula: see text].

Download Full-text

Identifying Model Weakness with Adversarial Examiner

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v34i07.6876 ◽

2020 ◽

Vol 34 (07) ◽

pp. 11998-12006 ◽

Cited By ~ 1

Author(s):

Michelle Shu ◽

Chenxi Liu ◽

Weichao Qiu ◽

Alan Yuille

Keyword(s):

Test Data ◽

Current Model ◽

Autonomous Driving ◽

Worst Case ◽

Average Case ◽

Constructive Feedback ◽

Data Space ◽

Systematic Exploration ◽

Testing Stage ◽

Performance Estimates

Machine learning models are usually evaluated according to the average case performance on the test set. However, this is not always ideal, because in some sensitive domains (e.g. autonomous driving), it is the worst case performance that matters more. In this paper, we are interested in systematic exploration of the input data space to identify the weakness of the model to be evaluated. We propose to use an adversarial examiner in the testing stage. Different from the existing strategy to always give the same (distribution of) test data, the adversarial examiner will dynamically select the next test data to hand out based on the testing history so far, with the goal being to undermine the model's performance. This sequence of test data not only helps us understand the current model, but also serves as constructive feedback to help improve the model in the next iteration. We conduct experiments on ShapeNet object classification. We show that our adversarial examiner can successfully put more emphasis on the weakness of the model, preventing performance estimates from being overly optimistic.

Download Full-text

Mosaic Mapping: A Means of Tiling Images to Shorten Acquisition Speed for Lower - Magnification SEM Images and Wavelength Dispersive Spectrometers (WDS) X-Ray Maps

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100164659 ◽

1996 ◽

Vol 54 ◽

pp. 438-439

Author(s):

J.D. Geller ◽

C.R. Herrington

Keyword(s):

Limiting Factors ◽

X Rays ◽

Angular Range ◽

Acceptance Angle ◽

Sem Images ◽

Worst Case ◽

X Ray ◽

Focusing Distance ◽

Layered Crystals ◽

Working Distance

The minimum magnification for which an image can be acquired is determined by the design and implementation of the electron optical column and the scanning and display electronics. It is also a function of the working distance and, possibly, the accelerating voltage. For secondary and backscattered electron images there are usually no other limiting factors. However, for x-ray maps there are further considerations. The energy-dispersive x-ray spectrometers (EDS) have a much larger solid angle of detection that for WDS. They also do not suffer from Bragg’s Law focusing effects which limit the angular range and focusing distance from the diffracting crystal. In practical terms EDS maps can be acquired at the lowest magnification of the SEM, assuming the collimator does not cutoff the x-ray signal. For WDS the focusing properties of the crystal limits the angular range of acceptance of the incident x-radiation. The range is dependent upon the 2d spacing of the crystal, with the acceptance angle increasing with 2d spacing. The natural line width of the x-ray also plays a role. For the metal layered crystals used to diffract soft x-rays, such as Be - O, the minimum magnification is approximately 100X. In the worst case, for the LEF crystal which diffracts Ti - Zn, ˜1000X is the minimum.

Download Full-text