Deep Learning

Discovering Symmetries in Neural Network Parameter Spaces
https://arxiv.org/abs/2609.33527

Augmented Lagrangian Predictive Coding
https://pub.sakana.ai/pc-alm/

DiffusionBlocks: Block-wise Neural Network Training via Diffusion Interpretation
https://arxiv.org/abs/2506.14202

Unified Neural Scaling Laws
https://arxiv.org/abs/2605.26248

Local learning for stable backpropagation-free neural network training towards physical learning
https://arxiv.org/abs/2603.24790

Topology and Geometry of the Learning Space of ReLU Networks: Connectivity and Singularities
https://arxiv.org/abs/2602.00693

Dense Neural Networks are not Universal Approximators
https://arxiv.org/abs/2602.07618

On neural scaling and the quanta hypothesis
https://ericjmichaud.com/quanta/

Interpret and manipulate the internals of deep learning models
https://github.com/ndif-team/nnsight

Revisiting Residual Connections: Orthogonal Updates for Stable and Efficient Deep Networks
https://arxiv.org/abs/2505.11881

Cautious Weight Decay
https://arxiv.org/abs/2510.12402

Pseudo-Invertible Neural Networks
https://yamitehr.github.io/spnn-project-page/

A Scalable Measure of Loss Landscape Curvature for Analyzing the Training Dynamics of LLMs
https://arxiv.org/abs/2601.16979

Understanding Warmup-Stable-Decay Learning Rates: A River Valley Loss Landscape Perspective
https://arxiv.org/abs/2410.05192

Nested Learning: The Illusion of Deep Learning Architectures
https://abehrouz.github.io/files/NL.pdf

Variational Continual Learning
https://arxiv.org/abs/1710.10628

On Surjectivity of Neural Networks: Can you elicit any behavior from your model?
https://arxiv.org/abs/2508.19445

Who Said Neural Networks Aren't Linear?
https://arxiv.org/abs/2510.08570

Hyperparameter Loss Surfaces Are Simple Near their Optima
https://arxiv.org/abs/2510.02721

Fisher Information Flow in Artificial Neural Networks
https://journals.aps.org/prx/abstract/10.1103/kn3z-rmm8

Approximate Controllability of Continuity Equation of Transformers
https://ieeexplore.ieee.org/abstract/document/10540044

Navigating the Latent Space Dynamics of Neural Models
https://arxiv.org/abs/2505.22785

Understanding Optimization in Deep Learning with Central Flows
https://arxiv.org/abs/2410.24206

Stochastic activations
https://arxiv.org/abs/2509.22358