# DANCE: Differentiable Accelerator/Network Co-Exploration

arXiv: https://arxiv.org/abs/2009.06237 · HTML: https://mlsanity.com/paper/2009.06237/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Sep 2020.

## Built on by (most cited first, 16 of 16)

- [A Comprehensive Survey on Hardware-Aware Neural Architecture Search](https://mlsanity.com/paper/2101.09336/) (2021)
- [Lightweight Deep Learning for Resource-Constrained Environments: A Survey](https://mlsanity.com/paper/2404.07236/) (2024)
- [EZNAS: Evolving Zero Cost Proxies For Neural Architecture Scoring](https://mlsanity.com/paper/2209.07413/) (2022)
- [Automated Deep Learning: Neural Architecture Search Is Not the End](https://mlsanity.com/paper/2112.09245/) (2021)
- [Rethinking Co-design of Neural Architectures and Hardware Accelerators](https://mlsanity.com/paper/2102.08619/) (2021)
- [CATransformers: Carbon Aware Transformers Through Joint Model-Hardware Optimization](https://mlsanity.com/paper/2505.01386/) (2025)
- [Special Session: Towards an Agile Design Methodology for Efficient, Reliable, and Secure ML Systems](https://mlsanity.com/paper/2204.09514/) (2022)
- [CODEBench: A Neural Architecture and Hardware Accelerator Co-Design Framework](https://mlsanity.com/paper/2212.03965/) (2022)
- [NAX: Co-Designing Neural Network and Hardware Architecture for Memristive Xbar based Computing Systems](https://mlsanity.com/paper/2106.12125/) (2021)
- [RHNAS: Realizable Hardware and Neural Architecture Search](https://mlsanity.com/paper/2106.09180/) (2021)
- [How to Reach Real-Time AI on Consumer Devices? Solutions for Programmable and Custom Architectures](https://mlsanity.com/paper/2106.15021/) (2021)
- [RT-RCG: Neural Network and Accelerator Search Towards Effective and Real-time ECG Reconstruction from Intracardiac Electrograms](https://mlsanity.com/paper/2111.02569/) (2021)
- [U-Boost NAS: Utilization-Boosted Differentiable Neural Architecture Search](https://mlsanity.com/paper/2203.12412/) (2022)
- [XploreNAS: Explore Adversarially Robust & Hardware-efficient Neural Architectures for Non-ideal Xbars](https://mlsanity.com/paper/2302.07769/) (2023)
- [Enabling Hard Constraints in Differentiable Neural Network and Accelerator Co-Exploration](https://mlsanity.com/paper/2301.09312/) (2023)
- [Uncertainty-Aware End-to-End Co-Design of Neural Network Processors: From Training and Mapping to Fabrication](https://mlsanity.com/paper/2606.04850/) (2026)

## Abstract (arXiv)

To cope with the ever-increasing computational demand of the DNN execution, recent neural architecture search (NAS) algorithms consider hardware cost metrics into account, such as GPU latency. To further pursue a fast, efficient execution, DNN-specialized hardware accelerators are being designed for multiple purposes, which far-exceeds the efficiency of the GPUs. However, those hardware-related metrics have been proven to exhibit non-linear relationships with the network architectures. Therefore it became a chicken-and-egg problem to optimize the network against the accelerator, or to optimize the accelerator against the network. In such circumstances, this work presents DANCE, a differentiable approach towards the co-exploration of the hardware accelerator and network architecture design. At the heart of DANCE is a differentiable evaluator network. By modeling the hardware evaluation software with a neural network, the relation between the accelerator architecture and the hardware metrics becomes differentiable, allowing the search to be performed with backpropagation. Compared to the naive existing approaches, our method performs co-exploration in a significantly shorter time, while achieving superior accuracy and hardware cost metrics.