# Efficacy of Neural Prediction-Based Zero-Shot NAS Source: [Efficacy of Neural Prediction-Based Zero-Shot NAS]() Collection index: [llms.txt]() Data updated: 2026-01-04T03:50:31.111Z Authors: Minh Le, Nhan Nguyen, Ngoc Hoang Luong Published: 2023-08-31T14:54:06.000Z ## Abstract In prediction-based Neural Architecture Search (NAS), performance indicators derived from graph convolutional networks have shown remarkable success. These indicators, achieved by representing feed-forward structures as component graphs through one-hot encoding, face a limitation: their inability to evaluate architecture performance across varying search spaces. In contrast, handcrafted performance indicators (zero-shot NAS), which use the same architecture with random initialization, can generalize across multiple search spaces. Addressing this limitation, we propose a novel approach for zero-shot NAS using deep learning. Our method employs Fourier sum of sines encoding for convolutional kernels, enabling the construction of a computational feed-forward graph with a structure similar to the architecture under evaluation. These encodings are learnable and offer a comprehensive view of the architecture's topological information. An accompanying multi-layer perceptron (MLP) then ranks these architectures based on their encodings. Experimental results show that our approach surpasses previous methods using graph convolutional networks in terms of correlation on the NAS-Bench-201 dataset and exhibits a higher convergence rate. Moreover, our extracted feature representation trained on each NAS benchmark is transferable to other NAS benchmarks, showing promising generalizability across multiple search spaces. The code is available at: https://github.com/minh1409/DFT-NPZS-NAS ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Architecture Performance Evaluation on DARTS search space CIFAR](): Architecture Performance Evaluation; DARTS search space CIFAR; Spearman Correlation: 0.758. - [Architecture Performance Evaluation on NASNet search space CIFAR](): Architecture Performance Evaluation; NASNet search space CIFAR; Spearman Correlation: 0.803. - [Architecture Performance Evaluation on Amoeba search space CIFAR](): Architecture Performance Evaluation; Amoeba search space CIFAR; Spearman Correlation: 0.797. - [Architecture Performance Evaluation on ENAS search space CIFAR](): Architecture Performance Evaluation; ENAS search space CIFAR; Spearman Correlation: 0.777. - [Architecture Performance Evaluation on PNAS search space CIFAR](): Architecture Performance Evaluation; PNAS search space CIFAR; Spearman Correlation: 0.758. - [Architecture Performance Evaluation on NAS-Bench-201 CIFAR](): Architecture Performance Evaluation; NAS-Bench-201 CIFAR; Spearman Correlation: 0.939. - [Architecture Performance Evaluation on NAS-Bench-101 CIFAR](): Architecture Performance Evaluation; NAS-Bench-101 CIFAR; Spearman Correlation: 0.877. - [Architecture Performance Evaluation on CIFAR Macro search space](): Architecture Performance Evaluation; CIFAR Macro search space; Spearman Correlation: 0.979. - [Score-Accuracy Correlation on NAS-Bench-201 (test)](): Score-accuracy correlation; NAS-Bench-201 (test); Spearman's Rho: 0.904. - [Image Classification on CIFAR-10 (test)](): Image Classification; CIFAR-10 (test); Accuracy: 96.7. - [Image Classification on CIFAR-100 (test)](): Image Classification; CIFAR-100 (test); Accuracy: 78.8. ## Original sources and implementation [Original paper]() [PDF]() [Implementation]()