Neural Networks
PulseAugur coverage of Neural Networks — every cluster mentioning Neural Networks across labs, papers, and developer communities, ranked by signal.
- instance of deep learning 90%
- used by gradient descent 90%
- instance of physics-informed neural networks 90%
- used by natural language processing 90%
- instance of Recurrent Neural Networks 90%
- used by deep learning 70%
- instance of gradient descent 70%
- competes with random forest 70%
- developed by physics-informed neural networks 70%
- uses Variational Inference 70%
- instance of computer vision 60%
- instance of random forest 50%
24 day(s) with sentiment data
-
New methods tackle data imbalance in regression tasks
Researchers have developed new methods to address data imbalance in regression tasks, a common issue that biases model performance, especially when predicting rare events. The study introduces novel sampling techniques,…
-
Qwen-AgentWorld paper explores new AI agent framework
A recent paper introduces Qwen-AgentWorld, a framework designed to enhance AI agent capabilities. The paper suggests this development aims to improve how AI agents interact and perform tasks, potentially leading to more…
-
New research explores neural network vulnerability to adversarial attacks · 4 sources tracked
Three new arXiv papers explore the complex relationship between neural network properties and adversarial robustness. One paper investigates over-parameterization, finding that while it enhances predictive capabilities,…
-
REViT: New Vision Transformer Achieves Roto-reflection Equivariance
Researchers have introduced REViT, a novel vision transformer that incorporates roto-reflection equivariance and convolutional attention. This approach aims to preserve rotational and flip symmetries in feature maps, wh…
-
New AI inference methods tackle high-dimensional variance and posterior collapse
Researchers have introduced Entropic Transport Descent (ETD), a novel particle-based variational inference method that uses entropy-regularized optimal transport to improve approximations of intractable distributions. U…
-
New inertial Dirac-Frenkel dynamics improve parameter stability in neural networks
Researchers have developed an inertial formulation of Dirac-Frenkel dynamics to address issues with non-unique or ill-conditioned parameter dynamics in redundant nonlinear parametrizations like neural networks. This new…
-
Noise fields used to spatially functionalize neural networks
Researchers have developed a novel approach to neural network computation by leveraging the spatial distribution of noise fields. This method, termed Spatial Partial Functionalization, utilizes a new activation function…
-
Neural networks reduce computational cost for vehicle aerodynamic simulations
Researchers have developed a new neural network-based approach for parametric model reduction to predict turbulent flow in vehicle aerodynamics. This method aims to reduce computational costs by projecting complex flow …
-
Local cycles identified as key design principle for neural network computation
Researchers have identified key structural design principles that enhance the computational abilities of recurrent neural networks. By training numerous networks to compute Boolean functions, they discovered that networ…
-
New SOAP-Bubbles method enhances neural network uncertainty estimation
Researchers have introduced SOAP-Bubbles, a novel method for estimating structured weight uncertainty in neural networks. This approach adapts the SOAP optimizer by running a variational method called IVON within the ei…
-
Introduction to Deep Learning and Neural Networks Explained
This article provides an introductory overview of deep learning and neural networks. It begins by exploring the biological neuron as the fundamental unit of human intelligence, laying the groundwork for understanding ar…
-
Biology's foundational role in AI development highlighted
The development of artificial intelligence, particularly neural networks, was significantly inspired by biological processes. Early artificial neurons mimicked the function of biological neurons, which receive, integrat…
-
Neural Networks Outperform Linear Regression in Complex Data Analysis
Neural networks offer significant advantages over linear regression, particularly in their ability to capture complex, non-linear patterns in data. They also possess self-organization and adaptability, allowing them to …
-
10 Core AI Concepts Explained: From Machine Learning to Robotics
This article provides a foundational overview of ten key concepts that underpin modern artificial intelligence. It aims to demystify complex AI topics by explaining core principles such as machine learning, deep learnin…
-
New synthetic data model aims to improve AI interpretability research
Researchers have introduced a novel synthetic data model called critical percolation, designed to better reflect the hierarchical structure found in natural data, which is often missing in current interpretability resea…
-
New SLiR method enhances neural network verification with broad applicability
Researchers have developed a new method called SLiR (Shifting-based Linear Relaxations) for verifying the behavior of neural networks. This approach is broadly applicable to various activation functions, requiring only …
-
New Adaptive Binning Method Enhances Tabular Self-Supervised Learning
Researchers have developed a new self-supervised learning technique called Adaptive Binning for tabular data, particularly in the medical field. This method improves upon existing approaches by adaptively refining featu…
-
Model-informed ML estimates carbon pools in European Shelf sea
Researchers have developed a model-informed machine learning approach to estimate carbon pools in the European Shelf sea environment. This method utilizes a deep ensemble of neural networks trained on observable variabl…
-
Neural networks offer new approach to neutrino mass ordering problem
Researchers have developed a novel machine-learning approach using neural networks to predict the neutrino mass ordering, a critical unsolved problem in particle physics. This method, trained on synthetic data from long…
-
Neural networks' loss of plasticity reduced by gradual environmental changes, study finds
A new research paper explores the phenomenon of "loss of plasticity" in neural networks, where the models gradually lose their ability to learn new tasks. The study, published on arXiv, investigates whether the abruptne…