Skip to search formSkip to main contentSkip to account menu

Lightweight Neural Architecture Search for Temporal Convolutional Networks at the Edge

This work proposes the first NAS tool that explicitly targets the optimization of the most peculiar architectural parameters of TCNs, namely dilation, receptive-field and number of features in each layer, and obtains superior solutions, while requiring low GPU memory and search time.
View on IEEE (opens in a new tab)

Robust and Energy-Efficient PPG-Based Heart-Rate Monitoring

This work proposes the use of hardware-friendly Temporal Convolutional Networks (TCN) for PPG-based heart estimation and obtains a TCN that outperforms the previous state-of-the- art on the largest PPG dataset available (PPGDalia).
View on IEEE (opens in a new tab)

Q-PPG: Energy-Efficient PPG-Based Heart Rate Monitoring on Wearable Devices

A design space exploration methodology to automatically generate a rich family of deep Temporal Convolutional Networks (TCNs) for HR monitoring, all derived from a single “seed” model, whose most accurate model sets a new state-of-the-art in Mean Absolute Error.
View on IEEE (opens in a new tab)

Pruning In Time (PIT): A Lightweight Network Architecture Optimizer for Temporal Convolutional Networks

An automatic dilation optimizer is proposed, which tackles the problem as a weight pruning on the time-axis, and learns dilation factors together with weights, in a single training, and reduces the model size and inference latency on a real SoC hardware target by up to 7.4× and 3×.
View on IEEE (opens in a new tab)

Embedding Temporal Convolutional Networks for Energy-efficient PPG-based Heart Rate Monitoring

This work derives a diverse set of Temporal Convolutional Networks for HR estimation, leveraging Neural Architecture Search, and introduces ActPPG, an adaptive algorithm that selects among multiple HR estimators depending on the amount of MAs, to improve energy efficiency.
View on ACM (opens in a new tab)

Deep Neural Network Architecture Search for Accurate Visual Pose Estimation aboard Nano-UAVs

This work deploys several NAS-optimized CNNs and runs them in closed-loop aboard a 27-g Crazyflie nano-UAV equipped with a parallel ultra-low power System-on-Chip, improving the State-of-the-Art by reducing the in-field control error and achieving a real-time onboard inference-rate of ~10Hz@10mW and ~50Hz@90mW.
View PDF on arXiv (opens in a new tab)

Channel-wise Mixed-precision Assignment for DNN Inference on Constrained Edge Nodes

This work proposes a novel NAS that selects the bit-width of each weight tensor channel independently, which gives the tool the additional flexibility of assigning a higher precision only to the weights associated with the most informative features.
View on IEEE (opens in a new tab)

Multi-Complexity-Loss DNAS for Energy-Efficient and Memory-Constrained Deep Neural Networks

Neural Architecture Search (NAS) is increasingly popular to automatically explore the accuracy versus computational complexity trade-off of Deep Learning (DL) architectures. When targeting tiny edge…
View on ACM (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)