Project Details

AppNeCo: Aproximativní neurovýpočty

Project Period: 1. 1. 2022 – 31. 12. 2024

Project Type: grant

Code: GA22-02067S

Agency: Czech Science Foundation

Program: Standardní projekty

English title
AppNeCo: Approximate Neurocomputing
Type
grant
Keywords

approximate computing,convolutional networks,energy complexity,robust learning,hardware accelerator,image classification

Abstract

Nowadays, modern AI technologies based on deep neural networks, whose computation is demanding on energy consumption, are implemented in devices with limited resources (e.g. battery powered cellphones). In error-tolerant applications (e.g. image classification), the use of approximate computing methods can save enormous amount of energy at the cost of only a small loss in accuracy. AppNeCo is a basic research project of approximate neurocomputing, whose ambition is an original synergy of approximation and complexity theory of neural networks and empirical experience with the top design of high-performance approximate implementations of hardware circuits. Its goal is to develop complexity-theoretic foundations of approximate computation by convolutional neural networks (CNN) of bounded energy complexity for application domains specified by input space distributions. This knowledge will be used in designing new strategies for approximating components and learning algorithms of low-energy high-precision CNNs. The new methods will be tested on image processing tasks.

Team members
Sekanina Lukáš, prof. Ing., Ph.D. (DCSY) – research leader
Klhůfek Jan, Ing. (DCSY)
Mrázek Vojtěch, Ing., Ph.D. (DCSY)
Vašíček Zdeněk, doc. Ing., Ph.D. (DCSY)
Publications

2024

  • MRÁZEK, V.; KOKKINIS, A.; PAPANIKOLAOU, P.; VAŠÍČEK, Z.; SIOZIOS, K.; TZIMPRAGOS, G.; TAHOORI, M.; ZERVAKIS, G. Evolutionary Approximation of Ternary Neurons for On-sensor Printed Neural Networks. 2024 IEEE/ACM International Conference on Computer Aided Design (ICCAD). New Jersey: 2024. p. 0-0. Detail
  • VAŠÍČEK, Z. Automated Synthesis of Commutative Approximate Arithmetic Operators. In 2024 IEEE Congress on Evolutionary Computation, CEC 2024 - Proceedings. Yokohama: IEEE Computer Society, 2024. p. 1-8. ISBN: 979-8-3503-0836-5. Detail

2023

  • KALKREUTH, R.; VAŠÍČEK, Z.; HUSA, J.; VERMETTEN, D.; YE, F.; THOMAS, B. General Boolean Function Benchmark Suite. In FOGA 2023 - Proceedings of the 17th ACM/SIGEVO Conference on Foundations of Genetic Algorithms. Potsdam: Association for Computing Machinery, 2023. p. 84-95. ISBN: 979-8-4007-0202-0. Detail
  • KALKREUTH, R.; VAŠÍČEK, Z.; HUSA, J.; VERMETTEN, D.; YE, F.; THOMAS, B. Towards a General Boolean Function Benchmark Suite. In GECCO 2023 Companion - Proceedings of the 2023 Genetic and Evolutionary Computation Conference Companion. New York: Association for Computing Machinery, 2023. p. 591-594. ISBN: 979-8-4007-0120-7. Detail
  • PIŇOS, M.; MRÁZEK, V.; SEKANINA, L. Prediction of Inference Energy on CNN Accelerators Supporting Approximate Circuits. In 2023 26th International Symposium on Design and Diagnostics of Electronic Circuits and Systems. Talinn: Institute of Electrical and Electronics Engineers, 2023. p. 45-50. ISBN: 979-8-3503-3277-3. Detail
  • SEDLÁČEK, M.; SEKANINA, L. Evolution of Editing Scripts From Examples. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (GECCO '23). Lisbon: Association for Computing Machinery, 2023. p. 803-806. ISBN: 979-8-4007-0120-7. Detail

2022

  • KLHŮFEK, J.; MRÁZEK, V. ArithsGen: Arithmetic Circuit Generator for Hardware Accelerators. In 2022 25th International Symposium on Design and Diagnostics of Electronic Circuits and Systems (DDECS '22). Prague: Institute of Electrical and Electronics Engineers, 2022. p. 44-47. ISBN: 978-1-6654-9431-1. Detail
  • KOCNOVÁ, J.; VAŠÍČEK, Z. Delay-aware evolutionary optimization of digital circuits. In Proceedings of IEEE Computer Society Annual Symposium on VLSI, ISVLSI. Nicosia, Cyprus: IEEE Computer Society, 2022. p. 188-193. ISBN: 978-1-6654-6605-9. Detail
  • MARCHISIO, A.; MRÁZEK, V.; MASSA, A.; BUSSOLINO, B.; MARTINA, M.; SHAFIQUE, M. RoHNAS: A Neural Architecture Search Framework with Conjoint Optimization for Adversarial Robustness and Hardware Efficiency of Convolutional and Capsule Networks. IEEE Access, 2022, vol. 2022, no. 10, p. 109043-109055. ISSN: 2169-3536. Detail
  • VÁLEK, M.; SEKANINA, L. Evolutionary Approximation in Non-Local Means Image Filters. In 2022 IEEE International Conference on Systems, Man, and Cybernetics (SMC). Praha: Institute of Electrical and Electronics Engineers, 2022. p. 2759-2766. ISBN: 978-1-6654-5258-8. Detail
Back to top