IEEE Design&Test Vol. 37, Issue 2

  • Speaker:
    Robust Machine Learning
  • Location:

    IEEE-Explorer

  • Date: March/April
Design & Test

Magazine
Volume 37, Issue 2 (March/April)

Highlights
Special Issue on "Robust Machine Learning"
Keynote by Sanjit A. Seshia, Somesh Jha, and Tommaso Dreossi "Semantic Adversarial Deep Learning "
Keynote by Malte J. Rasch, Tayfun Gokmen, and Wilfried Haensch "Training Large-scale Artificial Neural Networks on Simulated Resistive Crossbar Arrays"
General Interest Paper by Dennis R. E. Gnad, Jonas Krautter, Mehdi B. Tahoori, Falk Schellenberg, and Amir Moradi "Remote Electrical-level Security Threats to Multi-Tenant FPGAs"

March/April 2020 Content


From the EIC
Robust Machine Learning
  View full article (PDF).

Special Issue on Robust Resource-Constrained Systems for Machine Learning
Guest Editors’ Introduction: Robust Resource-Constrained Systems for Machine Learning
  View full article (PDF).
SSCNets: Robustifying DNNs using Secure Selective Convolutional Filters
  Training data is crucial in ensuring robust neural inference, and deep neural networks (DNNs) are heavily dependent on this assumption. However, DNNs can be exploited by adversaries that facilitate various attacks. read more
View full article (PDF).
Adaptive Neural Network Architectures for Power Aware Inference
  When dealing with edge devices, diverse power and compute constraints impose tradeoffs among performance, accuracy, and energy requirements during inference. read more.
View full article (PDF).
Are CNNs Reliable Enough for Critical Applications? An Exploratory Study
  Resource-constrained CNN implementations are subject to various reliability threats. read more.
View full article (PDF).
Impact of Memory Voltage Scaling on Accuracy and Resilience of Deep Learning Based Edge Devices
  As more and more artificial intelligence capabilities are deployed onto resource-constrained devices, designers explore several techniques in an effort to boost energy efficiency. read more
View full article (PDF).
Enabling Timing Error Resilience for Low-Power Systolic-Array Based Deep Learning Accelerators
  Hardware-accelerated learning and inference algorithms are quite popular in edge devices where predictable timing behavior and minimal energy consumption are required, while maintaining robustness to timing errors. read more
View full article (PDF).
Backdoor Suppression in Neural Networks using Input Fuzzing and Majority Voting
  While inference is needed at the edge, training is typically done at the cloud. Therefore, data necessary for training a model, as well as the trained model, have to be transmitted back and forth between the edge and the cloud training infrastructure. read more
View full article (PDF).

Survey Paper
Robust Machine Learning Systems: Challenges,Current Trends, Perspectives, and the Road Ahead
  Currently, machine learning (ML) techniques are at the heart of smart cyber-physical systems (CPS) and Internet-of-Things (IoT). read more
View full article (PDF).

Keynote Papers
Semantic Adversarial Deep Learning
  Adversarial examples have emerged as a key threat for machine-learning-based systems, especially the ones that employ deep neural networks. read more
View full article (PDF).
Training Large-scale Artificial Neural Networks on Simulated Resistive Crossbar Arrays
  Resistive crossbar arrays are promising options for accelerating enormous computation needed for training modern deep neural networks (DNNs). read more
View full article (PDF).

General Interest Paper
Remote Electrical-level Security Threats to Multi-Tenant FPGAs
  Virtualized FPGAs to provide multitenant access to increase their utilization have become a popular trend among the cloud computing providers. read more
View full article (PDF).

Departments
Report on the 38th ACM/IEEE International Conference on Computer-Aided Design (ICCAD 2019)
  View full article (PDF).
The Last Byte: Are You Sure You Love That Store?
  View full article (PDF).