# Runtime Configurable Deep Neural Networks for Energy-Accuracy Trade-off

arXiv: https://arxiv.org/abs/1607.05418 · HTML: https://mlsanity.com/paper/1607.05418/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Jul 2016.

## Built on by (most cited first, 19 of 19)

- [AppealNet: An Efficient and Highly-Accurate Edge/Cloud Collaborative Architecture for DNN Inference](https://mlsanity.com/paper/2105.04104/) (2021)
- [EcoFusion: Energy-Aware Adaptive Sensor Fusion for Efficient Autonomous Vehicle Perception](https://mlsanity.com/paper/2202.11330/) (2022)
- [DISTREAL: Distributed Resource-Aware Learning in Heterogeneous Systems](https://mlsanity.com/paper/2112.08761/) (2021)
- [Human Activity Recognition on Microcontrollers with Quantized and Adaptive Deep Neural Networks](https://mlsanity.com/paper/2209.00839/) (2022)
- [Optimising Resource Management for Embedded Machine Learning](https://mlsanity.com/paper/2105.03608/) (2021)
- [EBJR: Energy-Based Joint Reasoning for Adaptive Inference](https://mlsanity.com/paper/2110.10343/) (2021)
- [Adaptive Deep Learning for Efficient Visual Pose Estimation aboard Ultra-low-power Nano-drones](https://mlsanity.com/paper/2401.15236/) (2024)
- [AdaCon: Adaptive Context-Aware Object Detection for Resource-Constrained Embedded Devices](https://mlsanity.com/paper/2108.06850/) (2021)
- [Energy-Efficient Adaptive Machine Learning on IoT End-Nodes With Class-Dependent Confidence](https://mlsanity.com/paper/2204.03431/) (2022)
- [Adaptive Random Forests for Energy-Efficient Inference on Microcontrollers](https://mlsanity.com/paper/2205.13838/) (2022)
- [Dynamic Decision Tree Ensembles for Energy-Efficient Inference on IoT Edge Nodes](https://mlsanity.com/paper/2306.09789/) (2023)
- [Approximate Computing Survey, Part II: Application-Specific & Architectural Approximation Techniques and Applications](https://mlsanity.com/paper/2307.11128/) (2023)
- [Energy-Aware Deep Learning on Resource-Constrained Hardware](https://mlsanity.com/paper/2505.12523/) (2025)
- [Incremental Training and Group Convolution Pruning for Runtime DNN Performance Scaling on Heterogeneous Embedded Platforms](https://mlsanity.com/paper/2105.03600/) (2021)
- [Adaptive Test-Time Augmentation for Low-Power CPU](https://mlsanity.com/paper/2105.06183/) (2021)
- [Two-stage Human Activity Recognition on Microcontrollers with Decision Trees and CNNs](https://mlsanity.com/paper/2206.07652/) (2022)
- [A2P-MANN: Adaptive Attention Inference Hops Pruned Memory-Augmented Neural Networks](https://mlsanity.com/paper/2101.09693/) (2021)
- [Dynamic DNNs and Runtime Management for Efficient Inference on Mobile/Embedded Devices](https://mlsanity.com/paper/2401.08965/) (2024)
- [Do Vision Transformers Need All-to-All Attention? Global Communication Through Elastic Learned Cores](https://mlsanity.com/paper/2605.12491/) (2026)

## Abstract (arXiv)

We present a novel dynamic configuration technique for deep neural networks that permits step-wise energy-accuracy trade-offs during runtime. Our configuration technique adjusts the number of channels in the network dynamically depending on response time, power, and accuracy targets. To enable this dynamic configuration technique, we co-design a new training algorithm, where the network is incrementally trained such that the weights in channels trained in earlier steps are fixed. Our technique provides the flexibility of multiple networks while storing and utilizing one set of weights. We evaluate our techniques using both an ASIC-based hardware accelerator as well as a low-power embedded GPGPU and show that our approach leads to only a small or negligible loss in the final network accuracy. We analyze the performance of our proposed methodology using three well-known networks for MNIST, CIFAR-10, and SVHN datasets, and we show that we are able to achieve up to 95% energy reduction with less than 1% accuracy loss across the three benchmarks. In addition, compared to prior work on dynamic network reconfiguration, we show that our approach leads to approximately 50% savings in storage requirements, while achieving similar accuracy.