NAS-Bench-360: Benchmarking Neural Architecture Search on Diverse Tasks
Not on the map. Why?
Abstract (arXiv)
Most existing neural architecture search (NAS) benchmarks and algorithms prioritize well-studied tasks, e.g. image classification on CIFAR or ImageNet. This makes the performance of NAS approaches in more diverse areas poorly understood. In this paper, we present NAS-Bench-360, a benchmark suite to evaluate methods on domains beyond those traditionally studied in architecture search, and use it to address the following question: do state-of-the-art NAS methods perform well on diverse tasks? To construct the benchmark, we curate ten tasks spanning a diverse array of application domains, dataset sizes, problem dimensionalities, and learning objectives. Each task is carefully chosen to interoperate with modern CNN-based search methods while possibly being far-afield from its original development domain. To speed up and reduce the cost of NAS research, for two of the tasks we release the precomputed performance of 15,625 architectures comprising a standard CNN search space. Experimentally, we show the need for more robust NAS evaluation of the kind NAS-Bench-360 enables by showing that several modern NAS procedures perform inconsistently across the ten tasks, with many catastrophically poor results. We also demonstrate how NAS-Bench-360 and its associated precomputed results will enable future scientific discoveries by testing whether several recent hypotheses promoted in the NAS literature hold on diverse tasks. NAS-Bench-360 is hosted at https://nb360.ml.cmu.edu.
Key facts. Cited by 32 papers in the map (data through 7 Oct 2026). Builds on 40 of its 85 references that are in the map. First posted on arXiv 12 Oct 2021.
Builds on
40 of 85 references.
References · 40
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2016Densely Connected Convolutional Networkscited by 6,691
- 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
- 2016Wide Residual Networkscited by 3,057
- 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,806
- 2018DARTS: Differentiable Architecture Searchcited by 1,782
- 2017Learning Transferable Architectures for Scalable Image Recognitioncited by 1,097
- 2018Regularized Evolution for Image Classifier Architecture Searchcited by 902
- 2018Efficient Neural Architecture Search via Parameter Sharingcited by 831
- 2016Hyperband: A Novel Bandit-Based Approach to Hyperparameter Optimizationcited by 688
- 2018ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardwarecited by 666
- 2021Perceiver IO: A General Architecture for Structured Inputs & Outputscited by 616
28 more
- 2018BOHB: Robust and Efficient Hyperparameter Optimization at Scalecited by 466
- 2018Spherical CNNscited by 452
- 2020NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Searchcited by 412
- 2019NAS-Bench-101: Towards Reproducible Neural Architecture Searchcited by 340
- 2020FSD50K: An Open Dataset of Human-Labeled Sound Eventscited by 339
- 2019ROCKET: Exceptionally fast and accurate time series classification using random convolutional kernelscited by 336
- 2019Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examplescited by 312
- 2019PC-DARTS: Partial Channel Connections for Memory-Efficient Architecture Searchcited by 265
- 2019Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentationcited by 258
- 2019Searching for A Robust Neural Architecture in Four GPU Hourscited by 258
- 2019Random Search and Reproducibility for Neural Architecture Searchcited by 253
- 2020Neural Architecture Search without Trainingcited by 241
- 2018A System for Massively Parallel Hyperparameter Tuningcited by 201
- 2021Zero-Cost Proxies for Lightweight NAScited by 194
- 2019BANANAS: Bayesian Optimization with Neural Architectures for Neural Architecture Searchcited by 175
- 2021Neural Architecture Search on ImageNet in Four GPU Hours: A Theoretically Inspired Perspectivecited by 168
- 2020AutoML-Zero: Evolving Machine Learning Algorithms From Scratchcited by 137
- 2019One-Shot Neural Architecture Search via Self-Evaluated Template Networkcited by 69
- 2020NAS-Bench-1Shot1: Benchmarking and Dissecting One-shot Neural Architecture Searchcited by 64
- 2020NAS-Bench-NLP: Neural Architecture Search Benchmark for Natural Language Processingcited by 60
- 2021TransNAS-Bench-101: Improving Transferability and Generalizability of Cross-Task Neural Architecture Searchcited by 57
- 2019Densely Connected Search Space for More Flexible Neural Architecture Searchcited by 52
- 2018Deep Learning for Electromyographic Hand Gesture Signal Classification Using Transfer Learningcited by 46
- 2020HOLMES: Health OnLine Model Ensemble Serving for Deep Learning Models in Intensive Care Unitscited by 45
- 2020Geometry-Aware Gradient Algorithms for Neural Architecture Searchcited by 40
- 2021Rethinking Neural Operations for Diverse Taskscited by 18
- 2019deepCR: Cosmic Ray Rejection with Deep Learningcited by 11
- 2020sEMG Gesture Recognition with a Simple Model of Attentioncited by 10
Built on by
Citing papers per year: 2022: 3; 2023: 5; 2024: 12; 2025: 7; 2026: 5.
32 papers.
Sort:
By area:
- 2023Neural Architecture Search: Insights from 1000 Paperscited by 136
- 2025MLGym: A New Framework and Benchmark for Advancing AI Research Agentscited by 111
- 2025CodePDE: An Inference Framework for LLM-driven PDE Solver Generationcited by 37
- 2023Cross-Modal Fine-Tuning: Align then Refinecited by 34
- 2025Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interactioncited by 34
- 2024Tag-LLM: Repurposing General-Purpose LLMs for Specialized Domainscited by 30
- 2024UPS: Efficiently Building Foundation Models for PDE Solving via Cross-Modal Adaptationcited by 27
- 2022Efficient Architecture Search for Diverse Taskscited by 18
- 2025CAT: Content-Adaptive Image Tokenizationcited by 18
- 2024HW-GPT-Bench: Hardware-Aware Architecture Benchmark for Language Modelscited by 10
- 2024Learning Modality Knowledge Alignment for Cross-Modality Transfercited by 8
- 2024Robust NAS under adversarial training: benchmark, theory, and beyondcited by 6
20 more
- 2024einspace: Searching for Neural Architectures from Fundamental Operationscited by 5
- 2024Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generationcited by 5
- 2024Efficient Deep Learning Infrastructures for Embedded Computing Systems: A Comprehensive Survey and Future Envisioncited by 5
- 2026ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Responsecited by 5
- 2023On the Hyperparameter Loss Landscapes of Machine Learning Models: An Exploratory Studycited by 4
- 2022AutoML for Climate Change: A Call to Actioncited by 3
- 2022Automated Dominative Subspace Mining for Efficient Neural Architecture Searchcited by 3
- 2023DNArch: Learning Convolutional Neural Architectures by Backpropagationcited by 3
- 2023How Much Is Hidden in the NAS Benchmarks? Few-Shot Adaptation of a NAS Predictorcited by 2
- 2024What explains the success of cross-modal fine-tuning with ORCA?cited by 2
- 2025TusoAI: Agentic Optimization for Scientific Methodscited by 2
- 2026Rethinking Cross-Modal Fine-Tuning: Optimizing the Interaction Between Feature Alignment and Target Fittingcited by 1
- 2026Agentic Neural Architecture Searchcited by 1
- 2024MSTAR: Multi-Scale Backbone Architecture Search for Timeseries Classificationcited by 0
- 2024DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollerscited by 0
- 2024Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learningcited by 0
- 2025Variation Matters: from Mitigating to Embracing Zero-Shot NAS Ranking Function Variationcited by 0
- 2025confopt: A Library for Implementation and Evaluation of Gradient-based One-Shot NAS Methodscited by 0
- 2026YOLO-NAS-Bench: A Surrogate Benchmark with Self-Evolving Predictors for YOLO Architecture Searchcited by 0
- 2026NAS-Driven Hardware Accelerator Exploration for Edge AI and Quantization Effects on the Pareto Spacecited by 0
- 2026NAS-Driven Hardware Accelerator Exploration for Edge AI and Quantization Effects on the Pareto Space13 Aug 2026
- 2026Agentic Neural Architecture Search8 Jul 2026
- 2026ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Response23 Apr 2026
- 2026YOLO-NAS-Bench: A Surrogate Benchmark with Self-Evolving Predictors for YOLO Architecture Search10 Mar 2026
- 2026Rethinking Cross-Modal Fine-Tuning: Optimizing the Interaction Between Feature Alignment and Target Fitting26 Jan 2026
- 2025TusoAI: Agentic Optimization for Scientific Methods28 Sep 2025
- 2025confopt: A Library for Implementation and Evaluation of Gradient-based One-Shot NAS Methods22 Jul 2025
- 2025Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interaction9 Jun 2025
- 2025CodePDE: An Inference Framework for LLM-driven PDE Solver Generation13 May 2025
- 2025Variation Matters: from Mitigating to Embracing Zero-Shot NAS Ranking Function Variation27 Feb 2025
- 2025MLGym: A New Framework and Benchmark for Advancing AI Research Agents20 Feb 2025
- 2025CAT: Content-Adaptive Image Tokenization6 Jan 2025
20 more
- 2024Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning2 Dec 2024
- 2024Efficient Deep Learning Infrastructures for Embedded Computing Systems: A Comprehensive Survey and Future Envision3 Nov 2024
- 2024Learning Modality Knowledge Alignment for Cross-Modality Transfer27 Jun 2024
- 2024Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation13 Jun 2024
- 2024einspace: Searching for Neural Architectures from Fundamental Operations31 May 2024
- 2024HW-GPT-Bench: Hardware-Aware Architecture Benchmark for Language Models16 May 2024
- 2024What explains the success of cross-modal fine-tuning with ORCA?20 Mar 2024
- 2024Robust NAS under adversarial training: benchmark, theory, and beyond19 Mar 2024
- 2024DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollers14 Mar 2024
- 2024UPS: Efficiently Building Foundation Models for PDE Solving via Cross-Modal Adaptation11 Mar 2024
- 2024MSTAR: Multi-Scale Backbone Architecture Search for Timeseries Classification21 Feb 2024
- 2024Tag-LLM: Repurposing General-Purpose LLMs for Specialized Domains6 Feb 2024
- 2023How Much Is Hidden in the NAS Benchmarks? Few-Shot Adaptation of a NAS Predictor30 Nov 2023
- 2023On the Hyperparameter Loss Landscapes of Machine Learning Models: An Exploratory Study23 Nov 2023
- 2023Cross-Modal Fine-Tuning: Align then Refine11 Feb 2023
- 2023DNArch: Learning Convolutional Neural Architectures by Backpropagation10 Feb 2023
- 2023Neural Architecture Search: Insights from 1000 Papers20 Jan 2023
- 2022Automated Dominative Subspace Mining for Efficient Neural Architecture Search31 Oct 2022
- 2022AutoML for Climate Change: A Call to Action7 Oct 2022
- 2022Efficient Architecture Search for Diverse Tasks15 Apr 2022
- 2025MLGym: A New Framework and Benchmark for Advancing AI Research Agentscited by 111
- 2025TusoAI: Agentic Optimization for Scientific Methodscited by 2
- 2026Agentic Neural Architecture Searchcited by 1
- 2026Agentic Neural Architecture Search8 Jul 2026
- 2025TusoAI: Agentic Optimization for Scientific Methods28 Sep 2025
- 2025MLGym: A New Framework and Benchmark for Advancing AI Research Agents20 Feb 2025
- 2023Neural Architecture Search: Insights from 1000 Paperscited by 136
- 2025CodePDE: An Inference Framework for LLM-driven PDE Solver Generationcited by 37
- 2023Cross-Modal Fine-Tuning: Align then Refinecited by 34
- 2024Tag-LLM: Repurposing General-Purpose LLMs for Specialized Domainscited by 30
- 2024UPS: Efficiently Building Foundation Models for PDE Solving via Cross-Modal Adaptationcited by 27
- 2022Efficient Architecture Search for Diverse Taskscited by 18
- 2025CAT: Content-Adaptive Image Tokenizationcited by 18
- 2024HW-GPT-Bench: Hardware-Aware Architecture Benchmark for Language Modelscited by 10
- 2024Learning Modality Knowledge Alignment for Cross-Modality Transfercited by 8
- 2024Robust NAS under adversarial training: benchmark, theory, and beyondcited by 6
- 2024einspace: Searching for Neural Architectures from Fundamental Operationscited by 5
- 2024Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generationcited by 5
12 most cited of 28 All 28 (JSON)
- 2026NAS-Driven Hardware Accelerator Exploration for Edge AI and Quantization Effects on the Pareto Space13 Aug 2026
- 2026ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Response23 Apr 2026
- 2026YOLO-NAS-Bench: A Surrogate Benchmark with Self-Evolving Predictors for YOLO Architecture Search10 Mar 2026
- 2026Rethinking Cross-Modal Fine-Tuning: Optimizing the Interaction Between Feature Alignment and Target Fitting26 Jan 2026
- 2025confopt: A Library for Implementation and Evaluation of Gradient-based One-Shot NAS Methods22 Jul 2025
- 2025CodePDE: An Inference Framework for LLM-driven PDE Solver Generation13 May 2025
- 2025Variation Matters: from Mitigating to Embracing Zero-Shot NAS Ranking Function Variation27 Feb 2025
- 2025CAT: Content-Adaptive Image Tokenization6 Jan 2025
- 2024Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning2 Dec 2024
- 2024Efficient Deep Learning Infrastructures for Embedded Computing Systems: A Comprehensive Survey and Future Envision3 Nov 2024
- 2024Learning Modality Knowledge Alignment for Cross-Modality Transfer27 Jun 2024
- 2024Enhancing Cross-Modal Fine-Tuning with Gradually Intermediate Modality Generation13 Jun 2024
12 newest of 28 All 28 (JSON)