A5011164681- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Neural dynamics and brain function
- Neural Networks and Reservoir Computing
- Advanced Neural Network Applications
- Renal cell carcinoma treatment
- Cancer Genomics and Diagnostics
- Advanced Image and Video Retrieval Techniques
- Medical Image Segmentation Techniques
- Machine Learning and ELM
- Economic and Financial Impacts of Cancer
- Domain Adaptation and Few-Shot Learning
- Low-power high-performance VLSI design
- Renal and related cancers
- Medical Imaging Techniques and Applications
- Genetic factors in colorectal cancer
- Image and Object Detection Techniques
- Neuroscience and Neural Engineering
- Pulmonary Hypertension Research and Treatments
- Topic Modeling
- IoT and Edge/Fog Computing
- Colorectal Cancer Treatments and Studies
- Advanced MRI Techniques and Applications
- Neural Networks and Applications
- Water Quality Monitoring Technologies
East Suffolk and North Essex NHS Foundation Trust
2020-2024
National Health Service
2020-2024
Mid Essex Hospital Services NHS Trust
2017-2023
Purdue University West Lafayette
2016-2021
R.V. College of Engineering
2020-2021
Broomfield Hospital
2017-2020
D A Pandu Memorial RV Dental College and Hospital
2020
Bharathiar University
2019
Visvesvaraya Technological University
2008
National Institutes of Health
1997-2002
Spiking Neural Networks (SNNs) have recently emerged as a prominent neural computing paradigm. However, the typical shallow SNN architectures limited capacity for expressing complex representations, while training deep SNNs using input spikes has not been successful so far. Diverse methods proposed to get around this issue such converting off-line trained Artificial (ANNs) SNNs. ANN-SNN conversion scheme fails capture temporal dynamics of spiking system. On other hand, it is still difficult...
Spiking Neural Networks (SNNs) have recently attracted significant research interest as the third generation of artificial neural networks that can enable low-power event-driven data analytics. The best performing SNNs for image recognition tasks are obtained by converting a trained Analog Network (ANN), consisting Rectified Linear Units (ReLU), to SNN composed integrate-and-fire neurons with "proper" firing thresholds. converted typically incur loss in accuracy compared provided original...
Abstract Spiking Neural Networks (SNNs) have emerged as a powerful neuromorphic computing paradigm to carry out classification and recognition tasks. Nevertheless, the general purpose platforms custom hardware architectures implemented using standard CMOS technology, been unable rival power efficiency of human brain. Hence, there is need for novel nanoelectronic devices that can efficiently model neurons synapses constituting an SNN. In this work, we propose heterostructure composed Magnetic...
Spiking Neural Networks (SNNs) are fast becoming a promising candidate for brain-inspired neuromorphic computing because of their inherent power efficiency and impressive inference accuracy across several cognitive tasks such as image classification speech recognition. The recent efforts in SNNs have been focused on implementing deeper networks with multiple hidden layers to incorporate exponentially more difficult functional representations. In this paper, we propose pre-training scheme...
Spiking Neural Networks (SNNs) operate with asynchronous discrete events (or spikes) which can potentially lead to higher energy-efficiency in neuromorphic hardware implementations. Many works have shown that an SNN for inference be formed by copying the weights from a trained Artificial Network (ANN) and setting firing threshold each layer as maximum input received layer. These type of converted SNNs require large number time steps achieve competitive accuracy diminishes energy savings. The...
in the epidemiology and clinical association of sickle cell disease with malaria, bacterial viral infections (including SARS-CoV-2), suggests that should be included Integrated Management Childhood Illness programme to improve outcomes.Provision for diagnosis treatment incorporated into national health systems programming, an emphasis on delivering services primary care setting.COVID-19 is expected herald a global economic recession might result contraction international funding development...
Metastatic papillary renal cancer (PRC) has poor outcomes, and new treatments are required. There is a strong rationale for investigating mesenchymal epithelial transition receptor (MET) programmed cell death ligand-1 (PD-L1) inhibition in this disease. In study, the combination of savolitinib (MET inhibitor) durvalumab (PD-L1 investigated.This single-arm phase II trial explored (1,500 mg once every four weeks) (600 daily; ClinicalTrials.gov identifier: NCT02819596). Treatment-naïve or...
Spiking neural networks (SNNs) have emerged as a promising brain inspired neuromorphic-computing paradigm for cognitive system design due to their inherent event-driven processing capability. The fully connected (FC) shallow SNNs typically used pattern recognition require large number of trainable parameters achieve competitive classification accuracy. In this paper, we propose deep spiking convolutional network (SpiCNN) composed hierarchy stacked layers followed by spatial-pooling layer and...
Deep neural networks are biologically inspired class of algorithms that have recently demonstrated the state-of-the-art accuracy in large-scale classification and recognition tasks. Hardware acceleration deep is paramount importance to ensure their ubiquitous presence future computing platforms. Indeed, a major landmark enables efficient hardware accelerators for recent advances from machine learning community viability aggressively scaled binary networks. In this paper, we demonstrate how...
The efficiency of the human brain in performing classification tasks has attracted considerable research interest brain-inspired neuromorphic computing. Hardware implementations a system aims to mimic computations through interconnection neurons and synaptic weights. A leaky-integrate-fire (LIF) spiking model is widely used emulate dynamics neuronal action potentials. In this work, we propose spin based LIF neuron using magneto-electric (ME) switching ferro-magnets. voltage across ME oxide...
In this work, we propose ReStoCNet, a residual stochastic multilayer convolutional Spiking Neural Network (SNN) composed of binary kernels, to reduce the synaptic memory footprint and enhance computational efficiency SNNs for complex pattern recognition tasks. ReStoCNet consists an input layer followed by stacked layers hierarchical feature extraction, pooling dimensionality reduction, fully-connected inference. addition, introduce connections between improve learning capability deep SNNs....
Trees are used by animals, humans and machines to classify information make decisions. Natural tree structures displayed synapses of the brain involves potentiation depression capable branching is essential for survival learning. Demonstration such features in synthetic matter challenging due need host a complex energy landscape learning, memory electrical interrogation. We report experimental realization tree-like conductance states at room temperature strongly correlated perovskite...
Liquid state machine (LSM), a bio-inspired computing model consisting of the input sparsely connected to randomly interlinked reservoir (or liquid) spiking neurons followed by readout layer, finds utility in range applications varying from robot control and sequence generation action, speech, image recognition. LSMs stand out among other Recurrent Neural Network (RNN) architectures due their simplistic structure lower training complexity. Plethora recent efforts have been focused towards...
545 Background: Metastatic papillary renal cancer (PRC) has poor outcomes and there is need for new treatments. There a strong rationale investigating MET PD-L1 inhibition in this disease. In study, we investigate savolitinib (MET inhibitor) durvalumab (PD-L1 together. Methods: This single arm phase I/II trial explored at starting doses of 1500mg Q4W 600mg OD respectively, with 4wk run-in. Treatment naïve or previously treated patients metastatic PRC were included. Response rate (RR) (RECIST...
Neuromorphic algorithms are being increasingly deployed across the entire computing spectrum from data centers to mobile and wearable devices solve problems involving recognition, analytics, search inference. For example, large-scale artificial neural networks (popularly called deep learning) now represent state-of-the art in a wide ever-increasing range of video/image/audio/text recognition problems. However, growth sets network complexities have led learning becoming one most challenging...
Biologically-inspired spiking neural networks (SNNs) have attracted significant research interest due to their inherent computational efficiency in performing classification and recognition tasks. The conventional CMOS-based implementations of large-scale SNNs are power intensive. This is a consequence the fundamental mismatch between technology used realize neurons synapses, neuroscience mechanisms governing operation, leading area-expensive circuit designs. In this work, we present...
Deep neural networks (DNNs) have emerged as the state-of-the-art technique in a wide range of machine learning tasks for analytics and computer vision next generation embedded (mobile, IoT, wearable) devices. Despite their success, they suffer from high energy requirements. In recent years, inherent error resiliency DNNs has been exploited by introducing approximations at either algorithmic or hardware levels (individually) to obtain savings while incurring tolerable accuracy degradation....
In this work, we propose a Spiking Neural Network (SNN) consisting of input neurons sparsely connected by plastic synapses to randomly interlinked liquid, referred as Liquid-SNN, for unsupervised speech and image recognition. We adapt the strength interconnecting liquid using Spike Timing Dependent Plasticity (STDP), which enables self-learn general representation unique classes patterns. The presented learning methodology makes it possible infer class test directly neuronal spiking...
Brain-inspired learning models attempt to mimic the computations performed in neurons and synapses constituting human brain achieve its efficiency cognitive tasks. In this work, we propose Spike Timing Dependent Plasticity-based unsupervised feature using convolution-over-time Spiking Neural Network (SNN). We use shared weight kernels that are convolved with input patterns over time encode representative features, thereby improving sparsity as well robustness of model. show Convolutional SNN...
Multilayered artificial neural networks have found widespread utility in classification and recognition applications. The scale complexity of such together with the inadequacies general purpose computing platforms led to a significant interest development efficient hardware implementations. In this work, we focus on designing energy-efficient on-chip storage for synaptic weights, motivated primarily by observation that number synapses is orders magnitude larger than neurons. Typical digital...
In this work, we propose stochastic Binary Spiking Neural Network (sBSNN) composed of spiking neurons and binary synapses (stochastic only during training) that computes probabilistically with one-bit precision for power-efficient memory-compressed neuromorphic computing. We present an energy-efficient implementation the proposed sBSNN using `stochastic bit' as core computational primitive to realize synapses, which are fabricated in 90nm CMOS process, achieve efficient on-chip training...