Under the framework of graph-based learning, key to robust subspace clustering and learning is obtain a good similarity graph that eliminates effects errors retains only connections between data points from same (i.e., intrasubspace points). Recent works achieve performance by modeling into their objective functions remove inputs. However, these approaches face limitations structure should be known prior complex convex problem must solved. In this paper, we present novel method eliminate...

This paper introduces an event-driven feedforward categorization system, which takes data from a temporal contrast address event representation (AER) sensor. The proposed system extracts bio-inspired cortex-like features and discriminates different patterns using AER based tempotron classifier (a network of leaky integrate-and-fire spiking neurons). One the system's most appealing characteristics is its processing, with both input taking form events (spikes). was evaluated on posture dataset...

As a bio-inspired and emerging sensor, an event-based neuromorphic vision sensor has different working principle compared to the standard frame-based cameras, which leads promising properties of low energy consumption, latency, high dynamic range (HDR), temporal resolution. It poses paradigm shift sense perceive environment by capturing local pixel-level light intensity changes producing asynchronous event streams. Advanced technologies for visual sensing system autonomous vehicles from...

A lot of works have shown that frobenius-norm based representation (FNR) is competitive to sparse and nuclear-norm (NNR) in numerous tasks such as subspace clustering. Despite the success FNR experimental studies, less theoretical analysis provided understand its working mechanism. In this paper, we fill gap by building connections between NNR. More specially, prove that: 1) when dictionary can provide enough representative capacity, exactly NNR even though data set contains Gaussian noise,...

Spiking neural networks (SNNs) have received significant attention for their biological plausibility. SNNs theoretically at least the same computational power as traditional artificial (ANNs). They possess potential of achieving energy-efficient machine intelligence while keeping comparable performance to ANNs. However, it is still a big challenge train very deep SNN. In this brief, we propose an efficient approach build SNNs. Residual network (ResNet) considered state-of-the-art and...

Spiking neural networks (SNNs) are well-known as brain-inspired models with high computing efficiency, due to a key component that they utilize spikes information units, close the biological systems. Although spiking based energy efficient by taking advantage of discrete spike signals, their performance is limited current network structures and training methods. As typical SNNs cannot apply gradient descent rules directly into parameter adjustment artificial (ANNs). Aiming at this...

A new learning rule (Precise-Spike-Driven (PSD) Synaptic Plasticity) is proposed for processing and memorizing spatiotemporal patterns. PSD a supervised that analytically derived from the traditional Widrow-Hoff can be used to train neurons associate an input spike pattern with desired train. adaptation driven by error between actual output spikes, positive errors causing long-term potentiation negative depression. The amount of modification proportional eligibility trace triggered afferent...

Under the framework of spectral clustering, key subspace clustering is building a similarity graph, which describes neighborhood relations among data points. Some recent works build graph using sparse, low-rank, and l2 -norm-based representation, have achieved state-of-the-art performance. However, these methods suffered from following two limitations. First, time complexities are at least proportional to cube size, make those inefficient for solving large-scale problems. Second, they cannot...

Given a data set from union of multiple linear subspaces, robust subspace clustering algorithm fits each group points with low-dimensional and then clusters these even though they are grossly corrupted or sampled the dependent subspaces. Under framework spectral clustering, recent works using sparse representation, low rank representation their extensions achieve results by formulating errors (e.g., corruptions) into objective functions so that can be removed inputs. However, approaches have...

Spiking neural networks (SNNs) have manifested remarkable advantages in power consumption and event-driven property during the inference process. To take full advantage of low improve efficiency these models further, pruning methods been explored to find sparse SNNs without redundancy connections after training. However, parameter still hinders In human brain, rewiring process is highly dynamic, while synaptic maintain relatively brain development. Inspired by this, here we propose an...

Spiking Neural Networks (SNNs) have attracted significant attention for their energy-efficient and brain-inspired event-driven properties. Recent advancements, notably Spiking-YOLO, enabled SNNs to undertake advanced object detection tasks. Nevertheless, these methods often suffer from increased latency diminished accuracy, rendering them less suitable latency-sensitive mobile platforms. Additionally, the conversion of artificial neural networks (ANNs) frequently compromises integrity ANNs'...

Biological neurons use diverse temporal expressions of spikes to achieve efficient communication and modulation neural activities. Nonetheless, existing neuromorphic computing systems mainly simplified neuron models with limited spiking behaviors due high cost emulating these biological spike patterns. Here, we propose a compact reconfigurable design using the intrinsic dynamics NbO 2 -based unit excellent tunability in an electrochemical memory (ECRAM) emulate fast-slow bio-plausible...

This paper proposes a biologically plausible network architecture with spiking neurons for sequence recognition. is unified and consistent system functional parts of sensory encoding, learning, decoding. the first systematic model attempting to reveal neural mechanisms considering both upstream downstream together. The whole temporal framework, where precise timing spikes employed information processing cognitive computing. Experimental results show that competent perform recognition, being...

Inspired by biological evolution, a plethora of algorithms with evolutionary features have been proposed. These strengths in certain aspects, thus yielding better optimization performance particular problem. However, wide range problems, none them are superior to one another. Synergetic combination these is the potential ways ameliorate their search ability. Based on this idea, paper proposes an adaptive memetic computing as synergy genetic algorithm, differential and estimation distribution...

Spiking neural networks (SNNs) have received significant attention for their biological plausibility. SNNs theoretically at least the same computational power as traditional artificial (ANNs). They possess potential of achieving energy-efficiency while keeping comparable performance to deep (DNNs). However, it is still a big challenge train very SNN. In this paper, we propose an efficient approach build spiking version residual network (ResNet). ResNet considered kind state-of-the-art...

During the past few decades, remarkable progress has been made in solving pattern recognition problems using networks of spiking neurons. However, issue involving computational process from sensory encoding to synaptic learning remains underexplored, as most existing models or algorithms target only part process. Furthermore, many proposed literature neglect pay little attention information encoding, which makes them incompatible with neural-realistic signals encoded real-world stimuli. By...

Memory is a complex process across different brain regions and fundamental function for many cognitive behaviors. Emerging experimental results suggest that memories are represented by populations of neurons organized in categorical hierarchical manner. However, it still not clear how the neural mechanisms emulated computational models. In this paper, we present spatio-temporal memory (STM) model using spiking to explore formulation organization brain. Unlike previous approaches, employs...

The temporal credit assignment problem, which aims to discover the predictive features hidden in distracting background streams with delayed feedback, remains a core challenge biological and machine learning. To address this issue, we propose novel spatio-temporal algorithm called STCA for training deep spiking neural networks (DSNNs). We present new spatiotemporal error backpropagation policy by defining based loss function, is able network losses spatial domains simultaneously....

Spiking Neural Networks (SNNs) represent and transmit information in spikes, which is considered more biologically realistic computationally powerful than the traditional Artificial Networks. The spiking neurons encode useful temporal possess highly anti-noise property. feature extraction ability of typical SNNs limited by shallow structures. This paper focuses on improving virtue Convolutional (CNNs). CNNs can extract abstract features resorting to structure convolutional maps. We propose a...

Address event representation (AER) cameras have recently attracted more attention due to the advantages of high temporal resolution and low power consumption, compared with traditional frame-based cameras. Since AER record visual input as asynchronous discrete events, they are inherently suitable coordinate spiking neural network (SNN), which is biologically plausible energy-efficient on neuromorphic hardware. However, using SNN perform object classification still challenging, lack effective...

Event-based cameras have attracted increasing attention due to their advantages of biologically inspired paradigm and low power consumption. Since event-based record the visual input as asynchronous discrete events, they are inherently suitable cooperate with spiking neural network (SNN). Existing works SNNs for processing events mainly focus on task object recognition. However, from camera triggered by dynamic changes, which makes it an ideal choice capture actions in scene. Inspired dorsal...

Thanks to their event-driven nature, spiking neural networks (SNNs) are surmised be great computation-efficient models. The neurons encode beneficial temporal facts and possess excessive anti-noise properties. However, the high-quality encoding of spatio-temporal complexity also its training optimization SNNs restricted by means contemporary problem, this article proposes a novel hierarchical visual device explore how information transmits signifies in retina usage biologically manageable...