Tag: artificial neural networks
Browse our exclusive articles!
Simplicity Bias in Neural Networks: A Compression View
Explore how simplicity bias in neural networks is explained through compression and the Minimum Description Length principle for better AI robustness.
Robust Humanoid Locomotion via Joint Torque Perturbation
Enhance sim-to-real transfer for humanoid robots using joint torque space perturbation injection for improved locomotion robustness and adaptability.
Hamiltonian Dynamics and Bottleneck in Leaky ResNets
Explore Hamiltonian mechanics in Leaky ResNets revealing bottleneck structures and adaptive training for improved feature learning and efficiency.
Moonwalk: Memory-Efficient Gradient Computation Method
Discover Moonwalk, a novel technique that reduces memory use in deep learning by optimizing gradient computation without storing activations.
Enes Causal Discovery: Advanced Causal Inference Model
Discover how the Enes framework uses a mixture of experts to improve causal discovery in observational data, outperforming traditional methods.