-
Zhuyi: Perception Processing Rate Estimation for Safety in Autonomous Vehicles
Authors:
Yu-Shun Hsiao,
Siva Kumar Sastry Hari,
Michał Filipiuk,
Timothy Tsai,
Michael B. Sullivan,
Vijay Janapa Reddi,
Vasu Singh,
Stephen W. Keckler
Abstract:
The processing requirement of autonomous vehicles (AVs) for high-accuracy perception in complex scenarios can exceed the resources offered by the in-vehicle computer, degrading safety and comfort. This paper proposes a sensor frame processing rate (FPR) estimation model, Zhuyi, that quantifies the minimum safe FPR continuously in a driving scenario. Zhuyi can be employed post-deployment as an onli…
▽ More
The processing requirement of autonomous vehicles (AVs) for high-accuracy perception in complex scenarios can exceed the resources offered by the in-vehicle computer, degrading safety and comfort. This paper proposes a sensor frame processing rate (FPR) estimation model, Zhuyi, that quantifies the minimum safe FPR continuously in a driving scenario. Zhuyi can be employed post-deployment as an online safety check and to prioritize work. Experiments conducted using a multi-camera state-of-the-art industry AV system show that Zhuyi's estimated FPRs are conservative, yet the system can maintain safety by processing only 36% or fewer frames compared to a default 30-FPR system in the tested scenarios.
△ Less
Submitted 6 May, 2022;
originally announced May 2022.
-
Electronic transport descriptors for the rapid screening of thermoelectric materials
Authors:
Tianqi Deng,
Jose Recatala-Gomez,
Masato Ohnishi,
D. V. Maheshwar Repaka,
Pawan Kumar,
Ady Suwardi,
Anas Abutaha,
Iris Nandhakumar,
Kanishka Biswas,
Michael B. Sullivan,
Gang Wu,
Junichiro Shiomi,
Shuo-Wang Yang,
Kedar Hippalgaonkar
Abstract:
The discovery of novel materials for thermoelectric energy conversion has potential to be accelerated by data-driven screening combined with high-throughput calculations. One way to increase the efficacy of successfully choosing a candidate material is through its evaluation using transport descriptors. Using a data-driven screening, we selected 12 potential candidates in the trigonal ABX2 family,…
▽ More
The discovery of novel materials for thermoelectric energy conversion has potential to be accelerated by data-driven screening combined with high-throughput calculations. One way to increase the efficacy of successfully choosing a candidate material is through its evaluation using transport descriptors. Using a data-driven screening, we selected 12 potential candidates in the trigonal ABX2 family, followed by charge transport property simulations from first principles. The results suggest that carrier scattering processes in these materials are dominated by ionised impurities and polar optical phonons, contrary to the oft-assumed acoustic-phonon-dominated scattering. Combined with calculations of thermal conductivity based on three-phonon scattering, we predict p-type AgBiS2 and TlBiTe2 as potential high-performance thermoelectrics in the intermediate temperature range for low grade waste heat harvesting, with a predicted zT above 1 at 500 K. Using these data, we further derive ground-state transport descriptors for the carrier mobility and the thermoelectric power factor. In addition to low carrier mass, high dielectric constant was found to be an important factor towards high carrier mobility. A quadratic correlation between dielectric constant and transport performance was established and further validated with literature. Looking ahead, dielectric constant can potentially be exploited as an independent tuning knob for improving the thermoelectric performance.
△ Less
Submitted 9 January, 2021;
originally announced January 2021.
-
Making Convolutions Resilient via Algorithm-Based Error Detection Techniques
Authors:
Siva Kumar Sastry Hari,
Michael B. Sullivan,
Timothy Tsai,
Stephen W. Keckler
Abstract:
The ability of Convolutional Neural Networks (CNNs) to accurately process real-time telemetry has boosted their use in safety-critical and high-performance computing systems. As such systems require high levels of resilience to errors, CNNs must execute correctly in the presence of hardware faults. Full duplication provides the needed assurance but incurs a prohibitive 100% overhead. Algorithmic t…
▽ More
The ability of Convolutional Neural Networks (CNNs) to accurately process real-time telemetry has boosted their use in safety-critical and high-performance computing systems. As such systems require high levels of resilience to errors, CNNs must execute correctly in the presence of hardware faults. Full duplication provides the needed assurance but incurs a prohibitive 100% overhead. Algorithmic techniques are known to offer low-cost solutions, but the practical feasibility and performance of such techniques have never been studied for CNN deployment platforms (e.g., TensorFlow or TensorRT on GPUs). In this paper, we focus on algorithmically verifying Convolutions, which are the most resource-demanding operations in CNNs. We use checksums to verify convolutions, adding a small amount of redundancy, far less than full-duplication. We first identify the challenges that arise in employing Algorithm-Based Error Detection (ABED) for Convolutions in optimized inference platforms that fuse multiple network layers and use reduced-precision operations, and demonstrate how to overcome them. We propose and evaluate variations of ABED techniques that offer implementation complexity, runtime overhead, and coverage trade-offs. Results show that ABED can detect all transient hardware errors that might otherwise corrupt output and does so while incurring low runtime overheads (6-23%), offering at least 1.6X throughput to workloads compared to full duplication.
△ Less
Submitted 8 June, 2020;
originally announced June 2020.
-
HarDNN: Feature Map Vulnerability Evaluation in CNNs
Authors:
Abdulrahman Mahmoud,
Siva Kumar Sastry Hari,
Christopher W. Fletcher,
Sarita V. Adve,
Charbel Sakr,
Naresh Shanbhag,
Pavlo Molchanov,
Michael B. Sullivan,
Timothy Tsai,
Stephen W. Keckler
Abstract:
As Convolutional Neural Networks (CNNs) are increasingly being employed in safety-critical applications, it is important that they behave reliably in the face of hardware errors. Transient hardware errors may percolate undesirable state during execution, resulting in software-manifested errors which can adversely affect high-level decision making. This paper presents HarDNN, a software-directed ap…
▽ More
As Convolutional Neural Networks (CNNs) are increasingly being employed in safety-critical applications, it is important that they behave reliably in the face of hardware errors. Transient hardware errors may percolate undesirable state during execution, resulting in software-manifested errors which can adversely affect high-level decision making. This paper presents HarDNN, a software-directed approach to identify vulnerable computations during a CNN inference and selectively protect them based on their propensity towards corrupting the inference output in the presence of a hardware error. We show that HarDNN can accurately estimate relative vulnerability of a feature map (fmap) in CNNs using a statistical error injection campaign, and explore heuristics for fast vulnerability assessment. Based on these results, we analyze the tradeoff between error coverage and computational overhead that the system designers can use to employ selective protection. Results show that the improvement in resilience for the added computation is superlinear with HarDNN. For example, HarDNN improves SqueezeNet's resilience by 10x with just 30% additional computations.
△ Less
Submitted 25 February, 2020; v1 submitted 22 February, 2020;
originally announced February 2020.
-
ML-based Fault Injection for Autonomous Vehicles: A Case for Bayesian Fault Injection
Authors:
Saurabh Jha,
Subho S. Banerjee,
Timothy Tsai,
Siva K. S. Hari,
Michael B. Sullivan,
Zbigniew T. Kalbarczyk,
Stephen W. Keckler,
Ravishankar K. Iyer
Abstract:
The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault inj…
▽ More
The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault injection engine, which can mine situations and faults that maximally impact AV safety, as demonstrated on two industry-grade AV technology stacks (from NVIDIA and Baidu). For example, DriveFI found 561 safety-critical faults in less than 4 hours. In comparison, random injection experiments executed over several weeks could not find any safety-critical faults
△ Less
Submitted 1 July, 2019;
originally announced July 2019.
-
First-principles study of the thermoelectric properties of strained graphene nanoribbons
Authors:
Pei Shan Emmeline Yeo,
Michael B. Sullivan,
Kian ** Loh,
Chee Kwan Gan
Abstract:
We study the transport properties, in particular, the thermoelectric figure of merit ZT of armchair graphene nanoribbons, AGNR-N (for N=4-12, with widths ranging from 3.7 to 13.6~Å) through strain engineering, where N is the number of carbon dimer lines across the AGNR width. We find that the tensile strain applied to AGNR-$N$ changes the transport properties by modifying the electronic structures…
▽ More
We study the transport properties, in particular, the thermoelectric figure of merit ZT of armchair graphene nanoribbons, AGNR-N (for N=4-12, with widths ranging from 3.7 to 13.6~Å) through strain engineering, where N is the number of carbon dimer lines across the AGNR width. We find that the tensile strain applied to AGNR-$N$ changes the transport properties by modifying the electronic structures and phonon dispersion relations. The tensile strain increases the ZT value of the AGNR-$N$ families with N=3p and N=3p+2, where $p$ is an integer. Our analysis based on accurate density-functional theory calculations suggests a possible route to increase the ZT values of AGNR-$N$ for potential thermoelectric applications.
△ Less
Submitted 20 August, 2013;
originally announced August 2013.
-
First Principles Investigations of the Atomic, Electronic, and Thermoelectric Properties of Equilibrium and Strained Bi2Se3 & Bi2Te3, with van der Waals Interactions
Authors:
Xin Luo,
Michael B. Sullivan,
Su Ying Quek
Abstract:
Bi2Se3 and Bi2Te3 are layered compounds of technological importance, being excellent thermoelectric materials as well as topological insulators. We report density functional theory calculations of the atomic, electronic and thermoelectric properties of strained bulk and thin film Bi2Se3 and Bi2Te3, focusing on an appropriate description of van der Waals interactions. The calculations show that the…
▽ More
Bi2Se3 and Bi2Te3 are layered compounds of technological importance, being excellent thermoelectric materials as well as topological insulators. We report density functional theory calculations of the atomic, electronic and thermoelectric properties of strained bulk and thin film Bi2Se3 and Bi2Te3, focusing on an appropriate description of van der Waals interactions. The calculations show that the van der Waals Density functional with Cooper's exchange can reproduce closely the experimental interlayer distances in unstrained Bi2Se3 and Bi2Te3. Interestingly, we predict atomic structures that are in much better agreement with the experimentally determined structure from Nakajima than that obtained from Wyckoff, especially for Bi2Se3 where the difference in atomic structures qualitatively changes the electronic band structure. The band structure obtained using the Nakajima structure, and the theoretically optimized structure, are in much better agreement with previous reports of photoemission measurements, than that obtained using the Wyckoff structure. Fully optimizing atomic structures of bulk and thin film Bi2Se3 and Bi2Te3 under different in-plane and uniaxial strains, we predict that the electronic band gap of both the bulk materials and thin films decreases with tensile in-plane strain and increases with compressive in-plane strain. We also predict, using the semiclassical Boltzmann approach, that the magnitude of the n-type Seebeck coefficient of Bi2Te3 can be increased by the compressive in-plane strain, while that of Bi2Se3 can be increased with tensile in-plane strain. Further, the in-plane power factor of n-doped Bi2Se3 can be increased with compressive uniaxial strain, while that of n-doped Bi2Te3 can be increased by compressive in-plane strain. ...
△ Less
Submitted 7 August, 2013;
originally announced August 2013.
