Search for: All records

Abstract In this paper we present an adaptive synaptic array that can be used to improve the energy-efficiency of training machine learning (ML) systems. The synaptic array comprises of an ensemble of analog memory elements, each of which is a micro-scale dynamical system in its own right, storing information in its temporal state trajectory. The state trajectories are then modulated by a system level learning algorithm such that the ensemble trajectory is guided towards the optimal solution. We show that the extrinsic energy required for state trajectory modulation can be matched to the dynamics of neural network learning which leads to a significant reduction in energy-dissipated for memory updates during ML training. Thus, the proposed synapse array could have significant implications in addressing the energy-efficiency imbalance between the training and the inference phases observed in artificial intelligence (AI) systems. 

Pavement structures are designed to withstand continuous damage during their design life. Damage starts as soon as the pavement is open to traffic and increases with time. If maintenance activities are not considered in the initial design or considered but not applied during the service life, damage will grow to a point where rehabilitation may be the only and most expensive option left. In order to monitor the evolution of damage and its severity in pavement structures, a novel data compression approach based on cumulative measurements from a piezoelectric sensor is presented in this paper. Specifically, the piezoelectric sensor uses a thin film of polyvinylidene fluoride to sense the energy produced by the micro deformation generated due to the application of traffic loads. Epoxy solution has been used to encapsulate the membrane providing hardness and flexibility to withstand the high-loads and the high-temperatures during construction of the asphalt layer. The piezoelectric sensors have been exposed to three months of loading (approximately 1.0 million loads of 65 kN) at the French Institute of Science and Technology for Transport, Development and Networks (IFSTTAR) fatigue carrousel. Notably, the sensors survived the construction and testing. Reference measurements were made with a commercial conventional strain gauge specifically designed for measurements in hot mix asphalt layers. Results from the carrousel successfully demonstrate that the novel approach can be considered as a good indicator of damage progression, thus alleviating the need to measure strains in pavement for the purpose of damage tracking.