# Thales: Formulating and Estimating Architectural Vulnerability Factors for DNN Accelerators

arXiv: https://arxiv.org/abs/2212.02649 · HTML: https://mlsanity.com/paper/2212.02649/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 89 references that are in the map.
- First posted on arXiv 5 Dec 2022.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [YOLOv3: An Incremental Improvement](https://mlsanity.com/paper/1804.02767/) (2018)
- [VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection](https://mlsanity.com/paper/1711.06396/) (2017)
- [Pitfalls of Graph Neural Network Evaluation](https://mlsanity.com/paper/1811.05868/) (2018)
- [NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis](https://mlsanity.com/paper/1604.02808/) (2016)
- [Deep Speech: Scaling up end-to-end speech recognition](https://mlsanity.com/paper/1412.5567/) (2014)
- [TBT: Targeted Neural Network Attack with Bit Trojan](https://mlsanity.com/paper/1909.05193/) (2019)
- [Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation](https://mlsanity.com/paper/1804.06055/) (2018)
- [MFAS: Multimodal Fusion Architecture Search](https://mlsanity.com/paper/1903.06496/) (2019)
- [Terminal Brain Damage: Exposing the Graceless Degradation in Deep Neural Networks Under Hardware Fault Attacks](https://mlsanity.com/paper/1906.01017/) (2019)
- [Glimpse Clouds: Human Activity Recognition from Unstructured Feature Points](https://mlsanity.com/paper/1802.07898/) (2018)
- [Adversarial Defense Through Network Profiling Based Path Extraction](https://mlsanity.com/paper/1904.08089/) (2019)
- [TensorFI: A Flexible Fault Injection Framework for TensorFlow Applications](https://mlsanity.com/paper/2004.01743/) (2020)
- [Taking a Deeper Look at Pedestrians](https://mlsanity.com/paper/1501.05790/) (2015)
- [Ptolemy: Architecture Support for Robust Deep Learning](https://mlsanity.com/paper/2008.09954/) (2020)
- [Crescent: Taming Memory Irregularities for Accelerating Deep Point Cloud Analytics](https://mlsanity.com/paper/2204.10707/) (2022)
- [Tigris: Architecture and Algorithms for 3D Perception in Point Clouds](https://mlsanity.com/paper/1911.07841/) (2019)
- [RTNN: Accelerating Neighbor Search Using Hardware Ray Tracing](https://mlsanity.com/paper/2201.01366/) (2022)
- [CLEAR: Cross-Layer Exploration for Architecting Resilience - Combining Hardware and Software Techniques to Tolerate Soft Errors in Processor Cores](https://mlsanity.com/paper/1604.03062/) (2016)

## Built on by (most cited first, 2 of 2)

- [Evaluating and Enhancing Robustness of Deep Recommendation Systems Against Hardware Errors](https://mlsanity.com/paper/2307.10244/) (2023)
- [Characterizing Soft-Error Resiliency in Arm's Ethos-U55 Embedded Machine Learning Accelerator](https://mlsanity.com/paper/2404.09317/) (2024)

## Abstract (arXiv)

As Deep Neural Networks (DNNs) are increasingly deployed in safety critical and privacy sensitive applications such as autonomous driving and biometric authentication, it is critical to understand the fault-tolerance nature of DNNs. Prior work primarily focuses on metrics such as Failures In Time (FIT) rate and the Silent Data Corruption (SDC) rate, which quantify how often a device fails. Instead, this paper focuses on quantifying the DNN accuracy given that a transient error has occurred, which tells us how well a network behaves when a transient error occurs. We call this metric Resiliency Accuracy (RA). We show that existing RA formulation is fundamentally inaccurate, because it incorrectly assumes that software variables (model weights/activations) have equal faulty probability under hardware transient faults. We present an algorithm that captures the faulty probabilities of DNN variables under transient faults and, thus, provides correct RA estimations validated by hardware. To accelerate RA estimation, we reformulate RA calculation as a Monte Carlo integration problem, and solve it using importance sampling driven by DNN specific heuristics. Using our lightweight RA estimation method, we show that transient faults lead to far greater accuracy degradation than what todays DNN resiliency tools estimate. We show how our RA estimation tool can help design more resilient DNNs by integrating it with a Network Architecture Search framework.