You are here

Deep Neural Network Learning Based Tools for Embedded Systems Under Side Channel Attacks

Description:

TECHNOLOGY AREA(S): Electronics

OBJECTIVE: The Army Combat Capabilities Development Command (CCDC) Armament Center leads the Army for cyber secured weapons, sensors and systems. CCDC sponsored a series of new generations of embedded systems and communication systems development for weapons. The current efforts focus on the capabilities of using deep learning technologies to enhance both hardware and software in relevant dense urban environments. One key aspect of these efforts is to enhance weapons to defend against side-channel attacks (SCAs).

DESCRIPTION: The current efforts focus on the capabilities of using deep learning technologies to enhance both hardware and software in relevant dense urban environments. One key aspect of these efforts is to enhance weapons to defend against cyber attacks; AI/ML techniques to identify and counter SCAs are of particular interest under this SBIR [1-9].

PHASE I: Government expects that basic investigations can be accomplished during Phase I. While deep learning neural network outperformed existing approaches in SCAs, there are several standing questions that require further investigations. 1) What are the meanings of the activation functions and weights correspondent to the keys and architectures under SCAs? 2) How to extract the features or group of features correspondent to the different components in one system architecture? 3) How to assemble/refine a neural network if we have trained neural network models for general components (i.e. different type of memory architectures)? To fully understand and utilize this powerful technique, the offeror should: 1) Investigate the anatomy of the neural network. 2) Identify the neural network models for basic components in architectures. 3) Build and refine deep learning neural network using basic neural network models. 4) Compare TA, ML-based approaches with the proposed deep learning neural network. It is anticipated that the Phase I study will be unclassified.

PHASE II: Software/Hardware Implementations: during this phase, the Government expects the models/software modules developed in Phase I to be integrated into the existing sensors, weapons, and communication systems. We also expect the offeror to investigate plug-and-play hardware implementation that can upload the existing deep learning software. As an integrated component, this new hardware shall be inserted onto the existing sensors, weapons and communication system to perform real time cybersecurity. It is anticipated that this Phase will be executed at the SECRET level.

PHASE III: The government expect the offeror to provide software products based on deep learning SCAs, and hardware products with our deep learning software upload to perform real time guardiancies in cyber security for existing CCDC systems. These products will have military engineer/soldier friendly interfaces to assist training and reconfigurations thereof.

KEYWORDS: Deep Neural Network, Artificial Intelligence, Machine Learning, Hardware, Software

References:

L. Lerman, R. Poussier, G. Bontempi, O. Markowitch, and F. Standaert, “Template Attacks vs. Machine Learning Revisited (and the Curse of Dimensionality in Side-Channel Analysis),” in Constructive Side-Channel Analysis and Secure Design COSADE 2015, Berlin, Germany, 2015. Revised Selected Papers, 2015, pp. 20–33.; Chari, S., Rao, J.R., Rohatgi, P., “Template attacks.,” In: Kaliski Jr., B.S., Ko¸c, C¸.K., Paar, C. (eds.) CHES 2002. LNCS, vol. 2523, pp. 13–28. Springer, Heidelberg (2002); Schindler, W., Lemke, K., Paar, C. “A stochastic model for differential side channel cryptanalysis,” In: Rao, J.R., Sunar, B. (eds.) CHES 2005. LNCS, vol. 3659, pp. 30–46. Springer, Heidelberg (2005); G. Hospodar, B. Gierlichs, E. De Mulder, I. Verbauwhede, and J. Vandewalle, “Machine learning in side-channel analysis: a first study,” Journal of Cryptographic Engineering, vol. 1, pp. 293–302, 2011.; Alia Levina, Daria Sleptsova, Oleg Zaitsev, “Side-channel attacks and machine learning approach,” 2016 18th Conference of Open Innovations Association and Seminar on Information Security and Protection of Information Technology (FRUCT-ISPIT).; Liran Lerman, Gianluca Bontempi, and Olivier Markowitch, “Side channel attack:; B. Liu, K. Chen, M. Seo, J. Roveda, R. Lysecky. Evaluation of the Complexity of Automated Trace Alignment using Novel Power Obfuscation Methods, ACM Great Lakes Symposium on VLSI (GLSVLSI), 2018.; http://colah.github.io/posts/2015-08-Understanding-LSTMs/

US Flag An Official Website of the United States Government