Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Characterizing the Edge of Stability in Variational Training Without Priors

NeurIPS 2026 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Variational Bayesian deep learning

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Studies variational training of neural networks without a prior-based KL regularizer, characterizing edge-of-stability dynamics in that setting.

Problem

Variational deep learning requires a prior divergence term, but prior elicitation is hard, so practitioners often downweight the regularization term; the consequences of this practice are not well understood.

Approach

Analyzes variational training without priors (dropping or downweighting the KL term) and characterizes it through the edge-of-stability phenomenon. Abstract is truncated; further details are not available.

Closest papers by meaning

• Variational Deep Learning via Implicit Regularization (ICLR 2026 · similarity 0.92)
• Variational Learning Finds Flatter Solutions at the Edge of Stability (NeurIPS 2025 · similarity 0.90)
• Conflicting Biases at the Edge of Stability: Norm versus Sharpness Regularization (ICML 2026 · similarity 0.89)
• Loss Landscape Characterization of Neural Networks without Over-Parametrization (NeurIPS 2024 · similarity 0.87)
• Globally Convergent Variational Inference (NeurIPS 2024 · similarity 0.87)
• Grokking at the Edge of Numerical Stability (ICLR 2025 · similarity 0.87)