REVIEW 6 cited by
On the Pitfalls of Heteroscedastic Uncertainty Estimation with Probabilistic Neural Networks
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Capturing aleatoric uncertainty is a critical part of many machine learning systems. In deep learning, a common approach to this end is to train a neural network to estimate the parameters of a heteroscedastic Gaussian distribution by maximizing the logarithm of the likelihood function under the observed data. In this work, we examine this approach and identify potential hazards associated with the use of log-likelihood in conjunction with gradient-based optimizers. First, we present a synthetic example illustrating how this approach can lead to very poor but stable parameter estimates. Second, we identify the culprit to be the log-likelihood loss, along with certain conditions that exacerbate the issue. Third, we present an alternative formulation, termed $\beta$-NLL, in which each data point's contribution to the loss is weighted by the $\beta$-exponentiated variance estimate. We show that using an appropriate $\beta$ largely mitigates the issue in our illustrative example. Fourth, we evaluate this approach on a range of domains and tasks and show that it achieves considerable improvements and performs more robustly concerning hyperparameters, both in predictive RMSE and log-likelihood criteria.
Forward citations
Cited by 6 Pith papers
-
A Multi-Agent System for Motor Design Optimization via an FEA-AI Hybrid Approach
A multi-agent system with RAG, FEA, and uncertainty-driven AI surrogates automates IPMSM design optimization and outperforms pure FEA or AI baselines under fixed simulation budgets.
-
NeuroSymActive: Differentiable Neural-Symbolic Reasoning with Active Exploration for Knowledge Graph Question Answering
NeuroSymActive combines soft-unification symbolic modules, a neural path evaluator, and Monte-Carlo-style active exploration to reach strong answer accuracy on KGQA benchmarks while cutting graph lookups and model cal...
-
Mapping intratumoral heterogeneity through PET-derived washout and deep learning after proton therapy
A deep learning framework estimates voxel-wise biological washout parameters from simulated post-proton therapy PET, reducing errors by up to 60% and revealing intratumoral heterogeneity.
-
Controllable 3D Molecular Generation for Structure-Based Drug Design Through Bayesian Flow Networks and Gradient Integration
Gradient guidance inside Bayesian Flow Network updates generates 3D drug candidates with stronger predicted docking scores, better retrosynthesis feasibility, and improved kinase selectivity than diffusion baselines.
-
Principled Input-Output-Conditioned Post-Hoc Uncertainty Estimation for Regression Networks
IO-CUE trains a small auxiliary network on (input, frozen output) pairs with a detached Gaussian NLL objective to estimate prediction variance post-hoc, with augmented probe data improving OOD detection.
-
Uncertainty-Driven Radar-Inertial Fusion for Instantaneous 3D Ego-Velocity Estimation
A complex-valued neural network estimates 3D ego-velocity and uncertainty from raw radar data, fused with an IMU via EKF, reporting lower error than two baselines on ColoRadar.
Discussion (0). Continue with ORCID to comment.