Latest Articles for Neural Networks

NDOA algorithm offers improved accuracy in signal direction estimation.

Multi-task learning enables machines to improve performance by sharing knowledge across tasks.

New methods in forecasting enhance scientific predictions and efficiency.

GSANs improve data processing in complex structures like graphs and simplicial complexes.

SEED uses a selection of experts to improve learning over time.

Explore how the Hessian impacts machine learning model performance and training strategies.

Examining how deep neural networks learn and the challenges they face.

Discover how neural networks improve the efficiency of model predictive control.

Examining the complexities and strategies for neural network learning in various data types.

A flexible system for better training of large neural networks.

The study investigates universal neurons in GPT-2 models and their roles.

Momentum-SAM offers an efficient alternative to traditional training methods for neural networks.

A new method improves polar codes for memoryless and memory channels using neural networks.

Explore how interpretability illusions affect our view of neural networks.

NACHOS streamlines EENN design, enhancing efficiency and performance with automated tools.

Study reveals strong patterns in depthwise-separable CNNs linked to biological vision.

A new framework improves continual learning for tasks combining vision and language.

New method enhances neural networks against adversarial attacks using set-based inputs.

Research explores improving machine learning adaptability through meta-learning and Solomonoff Induction.

A new method for improving neural networks' resistance to attacks while maintaining performance.

A new method to create random samples from characteristic functions using neural networks.

RFMs improve feature learning and handle high-dimensional data efficiently.

This research focuses on enhancing few-shot learning through careful class selection.

A look at how input patterns affect stability in neural networks.

This research examines optimization guarantees of unfolded networks in machine learning.

New loss functions improve image classification in neural networks.

A look into the essential theory behind deep learning models.

SNNs offer energy-efficient solutions for NLP tasks compared to traditional models.

A straightforward look at how tensors interact in neural networks.

Discover the energy-efficient world of Spiking Neural Networks and their unique learning methods.

BlackMamba combines state-space models and mixture of experts for efficient language tasks.

Research highlights the importance of neuron interactions in sensory perception.

This study examines adding recurrence to Transformers for improved performance in machine learning tasks.

This method accelerates training for sequential models without sacrificing accuracy.

A look at how Transformers and GSSMs handle copying tasks.

Todyformer enhances dynamic graph analysis with efficient local and global learning.

Exploring the synergy between RL and LLMs for improved AI applications.

A new approach combines Gaussian components and mesh structures for efficient 3D rendering.

A look at bi-CryptoNets and their impact on data privacy.

Introducing a method to reduce forgetting in neural networks while learning new tasks.