A5037310533- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Neuroscience and Neural Engineering
- Neural dynamics and brain function
- Neural Networks and Reservoir Computing
- CCD and CMOS Imaging Sensors
- Neural Networks and Applications
- EEG and Brain-Computer Interfaces
- Memory and Neural Mechanisms
- Robotics and Sensor-Based Localization
- Photoreceptor and optogenetics research
- Robotic Path Planning Algorithms
- Advanced Neural Network Applications
- Semiconductor Lasers and Optical Devices
- GaN-based semiconductor devices and materials
- Recommender Systems and Techniques
- Organic Light-Emitting Diodes Research
- Distributed Control Multi-Agent Systems
- Advanced Graph Neural Networks
- Computability, Logic, AI Algorithms
- Privacy-Preserving Technologies in Data
Imec the Netherlands
2023-2025
Maastricht University
2024
Rutgers, The State University of New Jersey
2018-2022
National Chung Cheng University
2022
Rutgers Sexual and Reproductive Health and Rights
2020
Energy-efficient simultaneous localization and mapping (SLAM) is crucial for mobile robots exploring unknown environments. The mammalian brain solves SLAM via a network of specialized neurons, exhibiting asynchronous computations event-based communications, with very low energy consumption. We propose brain-inspired spiking neural (SNN) architecture that the unidimensional by introducing spike-based reference frame transformation, visual likelihood computation, Bayesian inference. integrated...
Energy-efficient mapless navigation is crucial for mobile robots as they explore unknown environments with limited on-board resources. Although the recent deep rein-forcement learning (DRL) approaches have been successfully applied to navigation, their high energy consumption limits use in several robotic applications. Here, we propose a neuromorphic approach that combines energy-efficiency of spiking neural networks optimality DRL and benchmark it control policies navigation. Our hybrid...
The field of neuromorphic computing holds great promise in terms advancing efficiency and capabilities by following brain-inspired principles. However, the rich diversity techniques employed research has resulted a lack clear standards for benchmarking, hindering effective evaluation advantages strengths methods compared to traditional deep-learning-based methods. This paper presents collaborative effort, bringing together members from academia industry, define benchmarks computing:...
The recent rise of Large Language Models (LLMs) has revolutionized the deep learning field. However, desire to deploy LLMs on edge devices introduces energy efficiency and latency challenges. Recurrent LLM (R-LLM) architectures have proven effective in mitigating quadratic complexity self-attention, making them a potential paradigm for computing on-edge neuromorphic processors. In this work, we propose low-cost, training-free algorithm sparsify R-LLMs' activations enhance hardware. Our...
Federated Learning (FL) allows collaboration between different parties, while ensuring that the data across these parties is not shared. However, every helpful in terms of resulting model performance. Therefore, it an important challenge to select correct participants a collaboration. As currently stands, most efforts participant selection literature have focused on Horizontal (HFL), which assumes all features are same participants, disregarding possibility captured Vertical (VFL). To close...
Locomotion is a crucial challenge for legged robots that addressed "effortlessly" by biological networks abundant in nature, named central pattern generators (CPG). The multitude of CPG network models have so far become biomimetic robotic controllers not applicable to the emerging neuromorphic hardware, depriving mobile robust walking mechanism would result inherently energy-efficient systems. Here, we propose brain-morphic controler based on comprehensive spiking neural-astrocytic generates...
Abstract Designing processors for implantable closed-loop neuromodulation systems presents a formidable challenge owing to the constrained operational environment, which requires low latency and high energy efficacy. Previous benchmarks have provided limited insights into power consumption latency. However, this study introduces algorithmic metrics that capture potential limitations of neural decoders intra-cortical brain–computer interfaces in context hardware constraints. This common...
The energy-efficient control of mobile robots is crucial as the complexity their real-world applications increasingly involves high-dimensional observation and action spaces, which cannot be offset by limited on-board resources. An emerging non-Von Neumann model intelligence, where spiking neural networks (SNNs) are run on neuromorphic processors, regarded an robust alternative to state-of-the-art real-time robotic controllers for low dimensional tasks. challenge now this new computing...
Neuromorphic processors aim to emulate the biological principles of brain achieve high efficiency with low power consumption. However, lack flexibility in most neuromorphic architecture designs results significant performance loss and inefficient memory usage when mapping various neural network algorithms. This paper proposes SENECA, a digital that balances trade-offs between using hierarchical-controlling system. A SENECA core contains two controllers, flexible controller (RISC-V) an...
The role of axonal synaptic delays in the efficacy and performance artificial neural networks has been largely unexplored. In step-based analog-valued network models (ANNs), concept is almost absent. their spiking neuroscience-inspired counterparts, there hardly a systematic account effects on model terms accuracy number operations. This paper proposes methodology for accounting training loop deep Spiking Neural Networks (SNNs), intending to efficiently solve machine learning tasks data with...
Sparse and event-driven spiking neural network (SNN) algorithms are the ideal candidate solution for energy-efficient edge computing. Yet, with growing complexity of SNN algorithms, it isn't easy to properly benchmark optimize their computational cost without hardware in loop. Although digital neuromorphic processors have been widely adopted black-box nature is problematic algorithm-hardware co-optimization. In this work, we open black box processor algorithm designers by presenting neuron...
Neuromorphic processors promise low-latency and energy-efficient processing by adopting novel brain-inspired design methodologies. Yet, current neuromorphic solutions still struggle to rival conventional deep learning accelerators' performance area efficiency in practical applications. Event-driven data-flow near/in-memory computing are the two dominant trends of processors. However, there remain challenges reducing overhead event-driven increasing mapping computing, which directly impacts...
It is true that the "best" neural network not necessarily one with most "brain-like" behavior. Understanding biological intelligence, however, a fundamental goal for several distinct disciplines. Translating our understanding of intelligence to machines problem in robotics. Propelled by new advancements Neuroscience, we developed spiking (SNN) draws from mounting experimental evidence number individual neurons associated spatial navigation. By following brain's structure, model assumes no...
Facilitated by the emergence of neuromorphic hardware, algorithms mimic brain's asynchronous computation to improve energy efficiency, low latency, and robustness, which are crucial for a wide variety real-time robotic applications. However, limited on-chip learning abilities hinder applicability computing real-world tasks. Biomimetism can overcome this limitation complementing or replacing training with knowledge connectome associated targeted behavior. By drawing inspiration from human...
Sparse and event-driven spiking neural network (SNN) algorithms are the ideal candidate solution for energy-efficient edge computing. Yet, with growing complexity of SNN algorithms, it isn't easy to properly benchmark optimize their computational cost without hardware in loop. Although digital neuromorphic processors have been widely adopted black-box nature is problematic algorithm-hardware co-optimization. In this work, we open black box processor algorithm designers by presenting neuron...
While there is still a lot to learn about astrocytes and their neuromodulatory role in the spatial temporal integration of neuronal activity, introduction neuromorphic hardware timely, facilitating computational exploration basic science questions as well exploitation real-world applications. Here, we present an astrocytic module that enables development spiking Neuronal-Astrocytic Network (SNAN) into Intel's Loihi chip. The basis end-to-end biophysically plausible compartmental model...
Energy-efficient mapless navigation is crucial for mobile robots as they explore unknown environments with limited on-board resources. Although the recent deep reinforcement learning (DRL) approaches have been successfully applied to navigation, their high energy consumption limits use in several robotic applications. Here, we propose a neuromorphic approach that combines energy-efficiency of spiking neural networks optimality DRL and benchmark it control policies navigation. Our hybrid...
While there is still a lot to learn about astrocytes and their neuromodulatory role in the spatial temporal integration of neuronal activity, introduction neuromorphic hardware timely, facilitating computational exploration basic science questions as well exploitation real-world applications. Here, we present an astrocytic module that enables development spiking Neuronal-Astrocytic Network (SNAN) into Intel's Loihi chip. The basis end-to-end biophysically plausible compartmental model...
Spiking neural networks (SNN) are delivering energy-efficient, massively parallel, and low-latency solutions to AI problems, facilitated by the emerging neuromorphic chips. To harness these computational benefits, SNN need be trained learning algorithms that adhere brain-inspired principles, namely event-based, local, online computations. Yet, state-of-the-art training based on backprop does not follow above principles. Due its limited biological plausibility, application of requires...
Neuromorphic processors are well-suited for efficiently handling sparse events from event-based cameras. However, they face significant challenges in the growth of computing demand and hardware costs as input resolution increases. This paper proposes Trainable Region-of-Interest Prediction (TRIP), first hardware-efficient hard attention framework vision processing on a neuromorphic processor. Our TRIP actively produces low-resolution (ROIs) efficient accurate classification. The exploits...
For Edge AI applications, deploying online learning and adaptation on resource-constrained embedded devices can deal with fast sensor-generated streams of data in changing environments. However, since maintaining low-latency power-efficient inference is paramount at the Edge, device should impose minimal additional overhead for inference. With this goal mind, we explore energy-efficient on-device streaming-data applications using Spiking Neural Networks (SNNs), which follow principles...
Spiking neural networks (SNNs) for event-based optical flow are claimed to be computationally more efficient than their artificial (ANNs) counterparts, but a fair comparison is missing in the literature. In this work, we propose an solution based on activation sparsification and neuromorphic processor, SENECA. SENECA has event-driven processing mechanism that can exploit sparsity ANN activations SNN spikes accelerate inference of both types networks. The have similar low activation/spike...
Currently, neural-network processing in machine learning applications relies on layer synchronization, whereby neurons a aggregate incoming currents from all the preceding layer, before evaluating their activation function. This is practiced even artificial Spiking Neural Networks (SNNs), which are touted as consistent with neurobiology, spite of brain being, fact asynchronous. A truly asynchronous system however would allow to evaluate concurrently threshold and emit spikes upon receiving...