Official implementation of Gradient-Update Alignment (GUA) for conflict-free training of Physics-Informed Neural Networks (PINNs).
-
Updated
Sep 18, 2026 - Python
#
gradient-conflict
Here are 8 public repositories matching this topic...
Open standard for processual awareness – machine-readable ontology of LifeNode
json-ld cybernetics resonance sheaf-theory ontology-engineering ascalon transduction chern-number-calculation gradient-conflict fiber-bundle timescape saddle-points bio-ai lifenode processual-inteligence processual-ontology machine-readable-ontology latent-frustration digital-mycelium bio-digital-coupling
-
Updated
Oct 21, 2026
通过基于“滞后梯度博弈”的内生演化压力，替代传统 MoE 的外在负载均衡约束，实现神经网络拓扑结构的自组织特化与层级化
pytorch deeplearning interpretability multi-task-learning mixture-of-experts conditional-computation disentanglement-learning gradient-conflict
-
Updated
Jan 12, 2026 - Python
[Electronics] Revisiting Dropout: Escaping Pressure for Training Neural Networks with Multiple Costs
-
Updated
Jul 19, 2021 - Python
Official PyTorch implementation of PRISM: Partition-of-Unity Routing for Isolated Subspace Manifolds in Dense Multi-Task Vision. Autonomous parameter-space AMR, continuous PoU gating, and vectorized gradient-clash decoupling for Vision Transformers.
computer-vision deep-learning pytorch semantic-segmentation nyuv2 depth-estimation multi-task-learning surface-normals adaptive-mesh-refinement mixture-of-experts scientific-machine-learning vision-transformers gradient-conflict partition-of-unity negative-transfer
-
Updated
Oct 11, 2026 - Python
How much of gradient conflict is the tasks, and how much the coordinates? DeepMath 2026 poster: code, results, project page.
-
Updated
Oct 25, 2026 - Python
Per-row gradient measurement at a language model's shared tied output projection: why cosine conflict diagnostics misread CE/MTP interference. Code, committed results, and the claim-verification gate for the preprint.
machine-learning deep-learning language-model multi-task-learning gradient-conflict multi-token-prediction
-
Updated
Aug 21, 2026 - Python
Official PyTorch implementation of ATLAS (Adaptive Tiling of Localized Anisotropic Subspaces) for autonomous parameter-space Adaptive Mesh Refinement (AMR) via vectorized gradient conflict profiling.
deep-learning reproducible-research quantum-mechanics pytorch computational-physics partial-differential-equations fluid-dynamics domain-decomposition adaptive-mesh-refinement scientific-machine-learning multiscale-modeling gradient-surgery physics-informed-neural-networks gradient-conflict partition-of-unity neural-operators surrogate-modeling spectral-bias fluid-dynamics-navier-stokes autonomous-amr
-
Updated
Oct 17, 2026 - Python
Add this topic to your repo
To associate your repository with the gradient-conflict topic, visit your repo's landing page and select "manage topics."