Understanding the structure and function of circuits is crucial for electronic design automation (EDA). Circuits can be formulated as And-Inverter graphs (AIGs), enabling efficient implementation representation learning through graph neural networks (GNNs). Masked modeling paradigms have been proven effective in learning. However, masking augmentation to original will destroy their logical equivalence, which unsuitable circuit Moreover, existing masked often prioritize structural information...

The photolithography process is getting more sophisticated with technology node scaling down and VLSI designs becoming complex. As photomask patterns get finer, mask rule checks (MRCs) are inevitable to avoid discrepancies in the layout ensure manufacturability. This paper introduces an efficient checking approach that utilizes a representative edge sampling scheme. scheme selects subset of edges points each polygon capture its contour, meanwhile greatly reducing number involved actual...

Running time is a key metric across the standard physical design flow stages. However, with rapid growth in sizes, routing runtime has become bottleneck flow. As result, speeding becomes critical and pressing task for IC automation. Aside from running time, we need to evaluate quality of global solution since poor engine degrades performance after entire stage. This work takes both them into consideration. We propose framework GPU-accelerated algorithms heterogeneous graph scheduler, called...

Deep reinforcement learning methods have shown great performance on many challenging cooperative multi-agent tasks. Two main promising research directions are value function decomposition and policy gradients. In this paper, we propose a new decomposed soft actor-critic (mSAC) method, which effectively combines the advantages of aforementioned two methods. The modules include Q network architecture, discrete probabilistic counterfactual advantage (optinal). Theoretically, mSAC supports...

A series of polyimides intrinsic microporosity (PIM-PIs) were prepared from commercially available rigid dianhydrides (pyromellitic dianhydride (PMDA), cyclo-[2.2.2]oct-7-ene-2-exo,3-exo,5-exo,6-exo-2,3:5,6-dianhydride (BTA), and 1,2,3,4-cyclobutanetetracarboxylic (CBDA)) diamines with norbornyl bis-benzocyclobutene (N2BC) segments (CANAL-2, CANAL-3, CANAL-4). The physical gas transport properties these polymers systematically investigated. fractional free volumes (FFV), interchain...

HVDC transmission, geomagnetic induction current, and subway DC stray current cause bias of the transformer, which leads to change transformer transmission characteristics affects accuracy relay protection action. Therefore, it is necessary analyze influence on itself. First, through a reasonable equivalent iron core, finite element model established based COMSOL software, comparison, verified that can truly reflect working state transformer. On this basis, used simulate changes secondary...

Building agents based on tree-search planning capabilities with learned models has achieved remarkable success in classic decision-making problems, such as Go and Atari. However, it been deemed challenging or even infeasible to extend Monte Carlo Tree Search (MCTS) algorithms diverse real-world applications, especially when these environments involve complex action spaces significant simulation costs, inherent stochasticity. In this work, we introduce LightZero, the first unified benchmark...

This paper proposes IncreMacro, a novel approach for macro placement refinement in the context of integrated circuit (IC) design. The suggested iteratively and incrementally optimizes macros order to enhance IC layout routability timing performance. To achieve this, IncreMacro utilizes several methods including kd-tree-based diagnosis, gradient-based shifting constraint-graph-based LP legalization. By employing these techniques iteratively, meets two critical solution requirements placement:...

A circuit design incorporating non-integer multi-height (NIMH) cells, such as a combination of 8-track and 12-track offers increased flexibility in optimizing area, timing, power simultaneously. The conventional approach for placing NIMH cells involves using commercial tools to generate an initial global placement, followed by legalization process that divides the block area into row regions with specific heights relocates rows matching height. However, placement flow often causes...

Fraudulent automobile insurance claims are not only a loss for companies, but also their policyholders. The goal of this research is to develop, first, decision-making algorithm classify whether claim classified as fraudulent or not; and, second, what types variables should be focused detect claims. To achieve goal, highly accurate prediction models built by discovering important sets features via variable selection algorithms, which can in turn help prevent future loss. In research,...

The performance of deep reinforcement learning methods prone to degenerate when applied tasks requiring relatively longer horizon memory or with highly variable dynamics. In this paper, we utilize the probabilistic latent context variables motivated by recent Meta-RL materials, and propose Latent Context based Soft Actor-Critic (LC-SAC) approach address aforementioned issues. is capable encode information about both agent's previous behaviors dynamics current undergoing environment, which...

BP neural network (Back-Propagation Neural Network, BP-NN) is one of the most widely models and applied to fault diagnosis power system currently. has good self-learning adaptive ability generalization ability, but operation process easy fall into local minima. Genetic algorithm global optimization features, crossover important Algorithm. In this paper, we can modify traditional Algorithm, using improved genetic optimized training initial weights thresholds, avoid problem The results...

The melt index(MI) is an important quality specification in polypropylene(PP) production industries. Accurate online measurement of MI becomes vital PP control. A new inferential model proposed by using Macro Heat Reaction(MHR) as a key factor correlation. MHR calculated from dynamic propylene polymerization model. validated plant data and the prediction error less than 5%. Simulations are performed three different models. results show that has least errors various operation conditions....

In order to meet the development of information exchange on a high speed radar tracking system, hybrid scheduling algorithm based CAN bus is presented. It combines static (DM, FPS) and dynamic (EDF, LLF) algorithm. Static has fixed priority can simply be used. However, its cannot changed, which means that deadline critical messages guaranteed. Dynamic efficiently improve bandwidth utilization guarantee transmission real-time messages. A introduced with CAN2.0B extended data frame protocol....

To address the problems that existing garment classification effects are greatly affected by background noise and network features not highly expressive. A multi-scale deep model (MCA-Inception) is proposed based on convolutional networks attention mechanisms. This uses modified Inception V3 as backbone expands perceptual field adding kernels of different scale sizes to enrich contextual detail information content. At same time, CBAM module embedded in improved suppress interference noisy...