This paper proposes Supervised Memory Training (SMT), a method for training nonlinear RNNs that sidesteps recurrent credit propagation entirely.

This paper introduces a framework to quantify information processing capacity in reservoir computing based on eigen-spectral decomposition, revealing vulnerabilities of low-energy modes.

The paper introduces an adaptive reservoir computing framework that tailors Echo State Networks (ESNs) to specific evaluation scenarios, achieving a high score on the CTF-4-Science Lorenz benchmark fo…

This paper studies the properties of random Boolean networks as reservoirs for reservoir computing and explores how design factors affect their computational capabilities.

This paper uses evolutionary optimization to improve reservoir computing for prediction tasks and reveals structural constraints on the recurrent substrate.

The paper proposes the Sector-Ring Linear Reservoir (SRLR) for reservoir computing, which improves performance by reducing the cost of search and adapting to different tasks.

The paper presents Flash RNN-T, a joint and loss implementation for efficient RNN-T training.

This paper presents AudioCALM, a universal audio generation framework that unifies speech, sound, and music generation.

The paper proposes CYKNN, a novel recurrent neural network architecture that directly encodes the CYK parsing algorithm, demonstrating superior performance over large language models on syntactic pars…

This paper introduces a mechanistic neuronal network model for multilayer learning, offering biological insights and an alternative to backpropagation.

The paper introduces LILAC, a codec idempotent neural audio codec that achieves competitive quality at 0.75 kbit/s and 9.375 Hz.

Echo is a joint-embedding predictive architecture that uses a single, pretrained ViT encoder to simultaneously perform speaker diarization, speech recognition, and dynamic source separation in a share…

This paper presents a novel interpretation of resonate-and-fire neural networks using state-space models, enabling parallel training with real and spiking inputs, and maintaining biologically-realisti…

A new error backpropagation method called supervised counterstream learning is proposed for deep associative networks, which only requires recognition of errors during training and backpropagates corr…