FPGA based spike-time dependent encoder and reservoir design in neuromorphic computing processors

Yi, Yang; Liao, Yulong; Wang, Bin; Fu, Xin; Shen, Fangyang; Hou, Hongyan; Liu, Lingjia

doi:10.1016/j.micpro.2016.03.009

Cited by 67 publications

(44 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Because of these features, neuromorphic computing is suitable for embedded systems. A lot of researchers [62][63][64][65][66] have implemented neuromorphic computing in embedded systems such as FPGA. In [63], the authors implement liquid state machine on FPGA for speech recognition.…”

Section: Neuromorphic Computing For Embedded Systemsmentioning

confidence: 99%

“…Researchers have successfully implemented neural encoder using hardware [10,62,69]. In [62], the authors proposed a spike time-dependent encoder on FPGA.…”

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

“…Researchers have successfully implemented neural encoder using hardware [10,62,69]. In [62], the authors proposed a spike time-dependent encoder on FPGA. In [69], the authors implemented an inter-spike interval-based encoder for neuromorphic processors using analog integrated circuits.…”

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

“…In recent years, an increasing number of researchers have started to implement neuromorphic computing using analog integrated circuits [46,47,49,50,71,[73][74][75][76][77][78][79]. Compared to digital implementation, analog implementation of neuromorphic computing is more energy efficient.…”

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

See 3 more Smart Citations

Quantized Neural Networks and Neuromorphic Computing for Embedded Systems

Liu¹,

Yi²

2020

Intelligent System and Computing

Self Cite

View full text Add to dashboard Cite

Deep learning techniques have made great success in areas such as computer vision, speech recognition and natural language processing. Those breakthroughs made by deep learning techniques are changing every aspect of our lives. However, deep learning techniques have not realized their full potential in embedded systems such as mobiles, vehicles etc. because the high performance of deep learning techniques comes at the cost of high computation resource and energy consumption. Therefore, it is very challenging to deploy deep learning models in embedded systems because such systems have very limited computation resources and power constraints. Extensive research on deploying deep learning techniques in embedded systems has been conducted and considerable progress has been made. In this book chapter, we are going to introduce two approaches. The first approach is model compression, which is one of the very popular approaches proposed in recent years. Another approach is neuromorphic computing, which is a novel computing system that mimicks the human brain.

show abstract

Section: Neuromorphic Computing For Embedded Systemsmentioning

confidence: 99%

“…Researchers have successfully implemented neural encoder using hardware [10,62,69]. In [62], the authors proposed a spike time-dependent encoder on FPGA.…”

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

Section: Recent Progress In Neuromorphic Computingmentioning

confidence: 99%

See 2 more Smart Citations

Quantized Neural Networks and Neuromorphic Computing for Embedded Systems

Liu¹,

Yi²

2020

Intelligent System and Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Such artificial intelligence is easy to train, e.g., using gradient descent machine learning algorithms --towards a specific task, and does not require an extensive technological or engineering overhead to implement [29]. As a result, reservoir computing has proven extremely useful for the design of a plethora of neuromorphic information processing applications that require fast real-time analysis of the time-series data [61,80,31,68,28].…”

Section: Computing Concretementioning

confidence: 99%

On Buildings that Compute. A Proposal

Adamatzky

Szaciłowski

Przyczyna

et al. 2019

From Astrophysics to Unconventional Computation

View full text Add to dashboard Cite

We present ideas aimed at bringing revolutionary changes on architectures and buildings of tomorrow by radically advancing the technology for the building material concrete and hence building components. We propose that by using nanotechnology we could embed computation and sensing directly into the material used for construction. Intelligent concrete blocks and panels advanced with stimuli-responsive smart paints are the core of the proposed architecture. In particular, the photo-responsive paint would sense the buildings internal and external environment while the nano-materialconcrete composite material would be capable of sensing the building environment and implement massive-parallel information processing resulting in distributed decision making. A calibration of the proposed materials with in-materio suitable computational methods and corresponding building in-formation modelling, computer-aided design and digital manufacturing tools could be achieved via models and prototypes of information processing at nano-level. The emergent technology sees a building as high-level massiveparallel computer -assembled of computing concrete blocks. Based on the generic principles of neuromorphic computation and reservoir computing we envisage a single building or an urban quarter to turn into a large-scale sensing substrate. It could behave as a universal computer, collecting and processing environmental information in situ enabling appropriate data fusion. The broad range of spatio-temporal effects include infrastructural and human mobility, energy, bio-diversity, digital activity, urban management, art and socializing, robustness with regard to damage and noise or real-time monitoring of environmental changes. The proposed intelligent architectures will increase sustainability and viability in digitised urban environments by decreasing information transfer bandwidth by e.g, utilising 5G networks. The emergence of socio-cultural effect will create a cybernetic relationship with our dwellings and cities.

show abstract

Electroformed, Self‐Connected Tin Oxide Nanoparticle Networks for Electronic Reservoir Computing

Murdoch

Barlow

et al. 2020

Adv Elect Materials

View full text Add to dashboard Cite

Interconnected SnO x nanoparticle networks have been electroformed within a semi-insulating SnO x thin-film and between lateral electrodes. During this top-down process, Joule heating, disproportionation and de-wetting of the SnO x thin-film preceded the formation of the nanoparticle networks. The same lateral electrodes used for electroforming were used to probe the network and reveal its complex electrical characteristics. Higher-order harmonic generation was observed and the internal short-term memory effects of the nanoparticle networks enabled temporal inputs to be mapped into reservoir states for subsequent linear readout without training. Reservoir computing functionality was demonstrated with no requirement for high-vacuum or cryo-cooled environments.

show abstract

FPGA based spike-time dependent encoder and reservoir design in neuromorphic computing processors

Cited by 67 publications

References 16 publications

Quantized Neural Networks and Neuromorphic Computing for Embedded Systems

Quantized Neural Networks and Neuromorphic Computing for Embedded Systems

On Buildings that Compute. A Proposal

Electroformed, Self‐Connected Tin Oxide Nanoparticle Networks for Electronic Reservoir Computing

Contact Info

Product

Resources

About