Generative design
Models that synthesize promising operators and architectures while respecting hardware constraints.
← All research fields
Research field 05
AI for Design Automation
How can AI become a trustworthy collaborator for hardware architects and EDA researchers?
Perspective
I use machine learning where it can expose structure in large design spaces: generating candidate circuits, learning surrogate models, guiding search, and balancing multiple objectives. The aim is not automation for its own sake, but faster and more explainable engineering insight.
Search & reinforcement learning
Tree search, constrained decoding, and learning-guided exploration across discrete design choices.
Surrogate & graph models
Fast prediction and relationship-aware reasoning for chiplets, reliability, and multi-objective trade-offs.
Relevant publications
Selected work in this field.
[DAC '26][ASP-DAC '23] [arXiv '25]
Yi-Cheng Lo and Siva Satyendra Sahoo, ChiSER: Surrogate-Enhanced Resource-aware Exploration of Chiplet-based Deep Neural Network Accelerators, in Proceedings of the 63rd ACM/IEEE Design Automation Conference (DAC '26), Long Beach, CA, USA, July 26-29, 2026, doi: 10.1145/3770743.3803969.
[DAC '26]Behnaz Ranjbar, Siva Satyendra Sahoo, Rohit Yalavarthy and Akash Kumar, XRel-Graph: Graph Learning-Driven Cross-Layer Reliability Management in Embedded Mixed-Criticality Systems, Accepted for DAC 2026.
[IEEE ESL '25]Ali Asghar, Shahzad Bangash, Suleman Shah, Laiq Hasan, Salim Ullah, Siva Satyendra Sahoo, and Akash Kumar, EMGAxO: Extending Machine Learning Hardware Generators With Approximate Operators, in IEEE Embedded Systems Letters, vol. 17, no. 5, pp. 345-348, Oct. 2025, doi: 10.1109/LES.2025.3600043.
[ACM TRETS '24]Siva Satyendra Sahoo, Salim Ullah, and Akash Kumar. 2024. AxOMaP: Designing FPGA-based Approximate Arithmetic Operators using Mathematical Programming. ACM Trans. Reconfigurable Technol. Syst. 17, 2, Article 31 (June 2024), 28 pages. https://doi.org/10.1145/3648694
[ACM TECS '23]Siva Satyendra Sahoo, Salim Ullah, and Akash Kumar. 2023. AxOTreeS: A Tree Search Approach to Synthesizing FPGA-based Approximate Operators. ACM Trans. Embed. Comput. Syst. 22, 5s, Article 101 (October 2023), 26 pages. https://doi.org/10.1145/3609096.
R. Ranjan, S. Ullah, S. S. Sahoo and A. Kumar, SyFAxO-GeN: Synthesizing FPGA-based Approximate Operators with Generative Networks, 2023 28th Asia and South Pacific Design Automation Conference (ASP-DAC ’23), Tokyo, Japan, 2023, pp. 402-409.
Sahoo, Siva Satyendra, Salim Ullah, and Akash Kumar. AxOSyn: An Open-source Framework for Synthesizing Novel Approximate Arithmetic Operators. ArXiv abs/2507.20007 (2025)