| Title | Attacking Split Manufacturing from a Deep Learning Perspective |
| Publication Type | Conference Paper |
| Year of Publication | 2019 |
| Authors | Li, H., Patnaik, S., Sengupta, A., Yang, H., Knechtel, J., Yu, B., Young, E. F. Y., Sinanoglu, O. |
| Conference Name | 2019 56th ACM/IEEE Design Automation Conference (DAC) |
| Date Published | jun |
| Keywords | back-end-of-line parts, BEOL connections, Capacitance, composability, Computer architecture, Deep Learning, different foundries, FEOL facility, Foundries, front-end-of-line, image processing, image-based features, industrial property, integrated circuit manufacture, integrated circuit split manufacturing, invasive software, IP piracy, ISCAS-85benchmarks, Layout, layout-level placement, learning (artificial intelligence), manufacturing systems, Metals, network-flow attack, neural nets, Pins, policy-based governance, pubcrawl, Resiliency, security of data, security promise, sophisticated deep neural network, vector-based features, Wires |
| Abstract | The notion of integrated circuit split manufacturing which delegates the front-end-of-line (FEOL) and back-end-of-line (BEOL) parts to different foundries, is to prevent overproduction, piracy of the intellectual property (IP), or targeted insertion of hardware Trojans by adversaries in the FEOL facility. In this work, we challenge the security promise of split manufacturing by formulating various layout-level placement and routing hints as vector- and image-based features. We construct a sophisticated deep neural network which can infer the missing BEOL connections with high accuracy. Compared with the publicly available network-flow attack [1], for the same set of ISCAS-85benchmarks, we achieve 1.21x accuracy when splitting on M1 and 1.12x accuracy when splitting on M3 with less than 1% running time. |
| Citation Key | li_attacking_2019 |
Attacking Split Manufacturing from a Deep Learning Perspective