The field of spiking neural networks (SNNs) and event-based vision is rapidly advancing, with a focus on developing more efficient and robust models. Recent research has explored the use of SNNs for binary classification in multivariate time series, with applications in areas such as gamma-ray spectral data and EEG recordings. Additionally, there is a growing interest in understanding the representational geometry of color qualia and developing more neurally plausible models. The use of event-based vision and tokenization techniques is also becoming increasingly popular, with applications in gesture recognition and object detection. Noteworthy papers in this area include: Exploring Spiking Neural Networks for Binary Classification in Multivariate Time Series at the Edge, which presents a general framework for training SNNs to perform binary classification on multivariate time series. Unveiling the Spatial-temporal Effective Receptive Fields of Spiking Neural Networks, which introduces the Spatio-Temporal Effective Receptive Field (ST-ERF) to analyze the ERF distributions across various Transformer-based SNNs. One-Timestep is Enough: Achieving High-performance ANN-to-SNN Conversion via Scale-and-Fire Neurons, which proposes a theoretical and practical framework for single-timestep ANN2SNN. Spiking Patches: Asynchronous, Sparse, and Efficient Tokens for Event Cameras, which presents a tokenizer specifically designed for event cameras.
Advances in Spiking Neural Networks and Event-Based Vision
Sources
Exploring Spiking Neural Networks for Binary Classification in Multivariate Time Series at the Edge
Seemingly Redundant Modules Enhance Robust Odor Learning in Fruit Flies
Unveiling the Spatial-temporal Effective Receptive Fields of Spiking Neural Networks
Probing the Representational Geometry of Color Qualia: Dissociating Pure Perception from Task Demands in Brains and AI Models
SARNet: A Spike-Aware consecutive validation Framework for Accurate Remaining Useful Life Prediction
One-Timestep is Enough: Achieving High-performance ANN-to-SNN Conversion via Scale-and-Fire Neurons
Conflict Adaptation in Vision-Language Models
Convolutional Spiking-based GRU Cell for Spatio-temporal Data
Spiking Patches: Asynchronous, Sparse, and Efficient Tokens for Event Cameras