Results 81 to 90 of about 112,191 (286)
Learning weakly supervised multimodal phoneme embeddings
, 2017 Recent works have explored deep architectures for learning multimodal speech representation (e.g. audio and images, articulation and audio) in a supervised way. Here we investigate the role of combining different speech modalities, i.e.
Chaabouni, Rahma +3 more
core +3 more sources
Neuromorphic Near‐Sensor and In‐Sensor Computing Enabled by Next‐Generation Material‐Based Sensors
Advanced Science, EarlyView.This Review presents a structural framework that classifies neuromorphic sensing into near‐sensor and in‐sensor architectures, clarifying physical coupling between sensing and computation. The framework connects neural and synaptic device functions with recent advances in optical, mechanical, and chemical sensing, compares energy consumption and ...
Su Yeon Jung +7 more
wiley +1 more source
DEKOMPOSISI TUGAS-TUGAS SOFTWARE-DEFINED RADIO (SDR) [PDF]
, 2009 ABTRACT This paper addresses decomposition of the tasks of Software-Defined Radio (SDR) computation by taking Gaussian Minimum Shift Keying (GMSK) as a case study.
Marpanaji, Eko
core
Photonic‐Enabled Energy‐Efficient Transparent Neuromorphic Computing Devices: A Review
Advanced Science, EarlyView.Transparent photonic neuromorphic computing devices merge optics and brain‐inspired computing to overcome von Neumann bottlenecks with ultrafast, low‐energy processing. By exploiting transparent oxides, 2D materials, phase‐change materials, and hybrid heterostructures, these platforms enable photonic synapses, memory, and logic for see‐through edge ...
Shuvaraj Ghosh +8 more
wiley +1 more source
RRAM Variability Harvesting for CIM‐Integrated TRNG
Advanced Electronic Materials, EarlyView.This work demonstrates a compute‐in‐memory‐compatible true random number generator that harvests intrinsic cycle‐to‐cycle variability from a 1T1R RRAM array. Parallel entropy extraction enables high‐throughput bit generation without dedicated circuits. This approach achieves NIST‐compliant randomness and low per‐bit energy, offering a scalable hardware
Ankit Bende +4 more
wiley +1 more source
International Journal of Reconfigurable Computing, 2018 CPU has insufficient resources to satisfy the efficient computation of the convolution neural network (CNN), especially for embedded applications. Therefore, heterogeneous computing platforms are widely used to accelerate CNN tasks, such as GPU, FPGA ...
Li Luo +9 more
doaj +1 more source
Task-Level Checkpointing System for Task-Based Parallel Workflows
, 2023 Scientific applications are large and complex; task-based programming models are a popular approach to developing these applications due to their ease of programming and ability to handle complex workflows and distribute their workload across large infrastructures.
Pere Vergés +3 more
openaire +2 more sources
Advanced Electronic Materials, EarlyView.Ising machines are emerging as specialized hardware solvers for computationally hard optimization problems. This review examines five major platforms—digital CMOS, analog CMOS, emerging devices, coherent optics, and quantum systems—highlighting physics‐rooted advantages and shared bottlenecks in scalability and connectivity.
Hyunjun Lee, Joon Pyo Kim, Sanghyeon Kim
wiley +1 more source
Omphale: Streamlining the Communication for Jobs in a Multi Processor System on Chip [PDF]
, 2007 Our Multi Processor System on Chip (MPSoC) template provides processing tiles that are connected via a network on chip. A processing tile contains a processing unit and a Scratch Pad Memory (SPM).
Bekooij, M.J.G. +3 more
core +2 more sources
Towards Task-Parallel Reductions in OpenMP
, 2015 Reductions represent a common algorithmic pattern in many scientific applications. OpenMP* has always supported them on parallel and worksharing constructs. OpenMP 3.0’s tasking constructs enable new parallelization opportunities through the annotation of irregular algorithms.
Jan Ciesko +10 more
openaire +3 more sources