HAS-RL: A Hierarchical Approximate Scheme Optimized With Reinforcement Learning for NoC-Based NN Accelerators

Network-on-Chip (NoC) is a scalable on-chip communication architecture for the NN accelerator, but with the increase in the number of nodes, the communication delay becomes higher. Applications such as machine learning have a certain resilience to noisy/erroneous transmitted data. Therefore, approximate communication becomes a promising solution...

Exploiting Symmetric Temporally Sparse BPTT for Efficient RNN Training

Recurrent Neural Networks (RNNs) are useful in temporal sequence tasks. However, training RNNs involves dense matrix multiplications which require hardware that can support a large number of arithmetic operations and memory accesses. Implementing online training of RNNs on the edge calls for optimized algorithms for an efficient deployment on...

Analysis of the mechanical properties of a fibreglass reinforced flexible pipe (FGRFP)

Fibreglass reinforced flexible pipe (FGRFP) is a kind of composite thermoplastic pipe serving as a preferred application in the field of oil transportation. This paper studies the mechanical behaviour of FGRFPs under pure bending by experimental, numerical and theoretical methods. Full-scale four-point bending tests are conducted and the...

A comparison of two fully coupled codes for integrated dynamic analysis of floating vertical axis wind turbines

This paper presents a comparison of two state-of-the-art codes that are capable of modelling floating vertical axis wind turbines (VAWTs) in fully coupled time-domain simulations, being the HAWC2 by DTU and the SIMO-RIFLEX-AC code by NTNU/MARINTEK. The comparative study focusses on the way aerodynamics, hydrodynamics and structural dynamics are...