#Neural Networks
Latest news articles tagged with "Neural Networks". Follow the timeline of events, related topics, and entities.
Articles (13)
-
🇺🇸 From Shallow Bayesian Neural Networks to Gaussian Processes: General Convergence, Identifiability and Scalable Inference
[USA]
arXiv:2602.22492v1 Announce Type: cross Abstract: In this work, we study scaling limits of shallow Bayesian neural networks (BNNs) via their connection to Gaussian processes (GPs), with an emphasis o...
Related: #Machine Learning Theory, #Statistical Modeling, #Computational Efficiency -
🇺🇸 Transformers converge to invariant algorithmic cores
[USA]
arXiv:2602.22600v1 Announce Type: cross Abstract: Large language models exhibit sophisticated capabilities, yet understanding how they work internally remains a central challenge. A fundamental obsta...
Related: #Machine Learning, #AI Interpretability -
🇺🇸 Elimination-compensation pruning for fully-connected neural networks
[USA]
arXiv:2602.20467v1 Announce Type: cross Abstract: The unmatched ability of Deep Neural Networks in capturing complex patterns in large and noisy datasets is often associated with their large hypothes...
Related: #Machine Learning, #Model Optimization -
🇺🇸 Learning During Detection: Continual Learning for Neural OFDM Receivers via DMRS
[USA]
arXiv:2602.20361v1 Announce Type: cross Abstract: Deep neural networks (DNNs) have been increasingly explored for receiver design because they can handle complex environments without relying on expli...
Related: #Machine Learning, #Wireless Communications -
🇺🇸 VINA: Variational Invertible Neural Architectures
[USA]
arXiv:2602.20480v1 Announce Type: cross Abstract: The distinctive architectural features of normalizing flows (NFs), notably bijectivity and tractable Jacobians, make them well-suited for generative ...
Related: #Machine Learning, #Theoretical Computer Science, #Scientific Research -
🇺🇸 Why Deep Jacobian Spectra Separate: Depth-Induced Scaling and Singular-Vector Alignment
[USA]
arXiv:2602.12384v1 Announce Type: cross Abstract: Understanding why gradient-based training in deep networks exhibits strong implicit bias remains challenging, in part because tractable singular-valu...
Related: #Deep Learning, #Mathematical Theory -
🇺🇸 Which Algorithms Can Graph Neural Networks Learn?
[USA]
arXiv:2602.13106v1 Announce Type: cross Abstract: In recent years, there has been growing interest in understanding neural architectures' ability to learn to execute discrete algorithms, a line of wo...
Related: #Algorithmic Reasoning, #Artificial Intelligence -
🇺🇸 Your Language Model Secretly Contains Personality Subnetworks
[USA]
arXiv:2602.07164v1 Announce Type: cross Abstract: Humans shift between different personas depending on social context. Large Language Models (LLMs) demonstrate a similar flexibility in adopting diffe...
Related: #Artificial Intelligence, #Machine Psychology -
🇺🇸 Free Energy Mixer
[USA]
arXiv:2602.07160v1 Announce Type: cross Abstract: Standard attention stores keys/values losslessly but reads them via a per-head convex average, blocking channel-wise selection. We propose the Free E...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 The Median is Easier than it Looks: Approximation with a Constant-Depth, Linear-Width ReLU Network
[USA]
arXiv:2602.07219v1 Announce Type: cross Abstract: We study the approximation of the median of $d$ inputs using ReLU neural networks. We present depth-width tradeoffs under several settings, culminati...
Related: #Machine Learning Theory, #Computational Complexity -
🇺🇸 Emergent Low-Rank Training Dynamics in MLPs with Smooth Activations
[USA]
arXiv:2602.06208v1 Announce Type: cross Abstract: Recent empirical evidence has demonstrated that the training dynamics of large-scale deep neural networks occur within low-dimensional subspaces. Whi...
Related: #Machine Learning, #Artificial Intelligence -
🇺🇸 Mining Generalizable Activation Functions
[USA]
arXiv:2602.05688v1 Announce Type: cross Abstract: The choice of activation function is an active area of research, with different proposals aimed at improving optimization, while maintaining expressi...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 Understanding neural networks through sparse circuits
[USA]
OpenAI is exploring mechanistic interpretability to understand how neural networks reason. Our new sparse model approach could make AI systems more transparent and support safer, more reliable behavio...
Related: #AI Transparency, #Safety and Reliability