Pith. sign in

REVIEW 3 cited by

An Uncertainty-aware Loss Function for Training Neural Networks with Calibrated Predictions

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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

arxiv 2110.03260 v2 pith:JXAFTV2B submitted 2021-10-07 cs.LG cs.CV

Afshar Shamsi , Hamzeh Asgharnezhad , AmirReza Tajally , Saeid Nahavandi , Henry Leung This is my paper
classification cs.LGcs.CV
keywords lossuncertaintyfunctionslearningmethodscalibrateddeepentropy
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Uncertainty quantification of machine learning and deep learning methods plays an important role in enhancing trust to the obtained result. In recent years, a numerous number of uncertainty quantification methods have been introduced. Monte Carlo dropout (MC-Dropout) is one of the most well-known techniques to quantify uncertainty in deep learning methods. In this study, we propose two new loss functions by combining cross entropy with Expected Calibration Error (ECE) and Predictive Entropy (PE). The obtained results clearly show that the new proposed loss functions lead to having a calibrated MC-Dropout method. Our results confirmed the great impact of the new hybrid loss functions for minimising the overlap between the distributions of uncertainty estimates for correct and incorrect predictions without sacrificing the model's overall performance.

Discussion (0). Sign in to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Uncertainty Estimation by Human Perception versus Neural Models

    cs.LG 2025-06 conditional novelty 5.0 of 10

    Neural network uncertainty estimates correlate only weakly with human-perceived uncertainty on three vision benchmarks, and soft-label training improves that alignment, though the claimed calibration benefit is not me...

  2. CLUE: Neural Networks Calibration via Learning Uncertainty-Error alignment

    cs.LG 2025-05 conditional novelty 4.0 of 10

    CLUE trains networks with a loss that penalizes mismatch between predicted uncertainty and the model's loss, claiming better calibration across vision, regression, and language tasks.

  3. Enhancing Monte Carlo Dropout Performance for Uncertainty Quantification

    cs.CV 2025-05 reject novelty 3.0 of 10

    Tuning Monte Carlo Dropout hyperparameters with GWO, BO, or PSO and adding a predictive-entropy loss term reportedly improves accuracy, uncertainty accuracy, and calibration by 2-3% over vanilla MCD.

Pith tools