Automatic transistor sizing is a challenging problem in circuit design due to the large design space, complex performance tradeoffs, and fast technology advancements. Although there have been plenty of work on transistor sizing targeting on one circuit, limited research has been done on transferring the knowledge from one circuit to another to reduce the re-design overhead. In this paper, we present GCN-RL Circuit Designer, leveraging reinforcement learning (RL) to transfer the knowledge between different technology nodes and topologies. Moreover, inspired by the simple fact that circuit is a graph, we learn on the circuit topology representation with graph convolutional neural networks (GCN). The GCN-RL agent extracts features of the topology graph whose vertices are transistors, edges are wires. Our learning-based optimization consistently achieves the highest Figures of Merit (FoM) on four different circuits compared with conventional black box optimization methods (Bayesian Optimization, Evolutionary Algorithms), random search and human expert designs. Experiments on transfer learning between five technology nodes and two circuit topologies demonstrate that RL with transfer learning can achieve much higher FoMs than methods without knowledge transfer. Our transferable optimization method makes transistor sizing and design porting more effective and efficient.
Li J, Gao X, Jiang T. Graph Networks for Multiple Object Tracking, in IEEE Winter Conference on Applications of Computer Vision, WACV 2020, Snowmass Village, CO, USA, March 1-5, 2020. IEEE; 2020:708–717. 访问链接
Zeng L, Zou L, Özsu TM, Hu L, Zhang F. GSI: GPU-friendly Subgraph Isomorphism, in 36th IEEE International Conference on Data Engineering, ICDE 2020, Dallas, TX, USA, April 20-24, 2020. IEEE; 2020:1249–1260.
Song T (PhD student), T Q, Chen J *. Multi task based sound localization model, in Proceedings of 148th Audio Engineering Society International Convention. Vienna, Austria; 2020:1-6.
Li X, Hu S, Zou L. Natural Answer Generation via Graph Transformer, in Web and Big Data - 4th International Joint Conference, APWeb-WAIM 2020, Tianjin, China, September 18-20, 2020, Proceedings, Part I.Vol 12317. Springer; 2020:302–318.
This paper discusses an accepted emendation to an earlier version of IG X 2.1 137. Early draft copies of the Herennia announcement show that Antoninus Pius was hailed as Σωτήρ by the city of Thessalonike, a rare epithet for this emperor. This reading was later replaced due to an expert's claim that σωτῆρος was σωτηρίας. Since this seems to conform to a well-known salutary formula, the emendation was adopted from then on. This paper wishes to suggest that the reading of σωτῆρος is based on reliable and published reports instead, and ought to be preferred over the expert claim. Empirical evidence is given to support reading σωτῆρος.