Fully Bayesian learning and classic deep learning: An efficient unified method for estimating uncertainty
Bibliographic record
Abstract
Classic deep learning algorithms are powerful tools for the construction of accurate predictive models for labeled data.However, traditional deep neural networks designed to learning such models are both prone to overfitting and incapable of assessing uncertainty.In contrast, Bayesian learning based upon the emergence of Markov chain Monte Carlo methods and variational inference provides strong ability to express uncertainty in predictions and improve the estimated posterior probability based on new evidence.This work further assesses the efficiency and accuracy of Bayesian inference in complex settings.We provide an in-depth empirical analysis of the methods on both real and synthetic data in the context of regression and image classification.Specifically, we develop a unified Bayesian deep neural network model interleaving Bayesian sampling into deep learning.By rephrasing these learning techniques upon a common theoretical ground casting (1) the application of fully Bayesian learning for deep neural networks rather than pure optimization-based or approximate learning and (2) the most significant regularization technique in neural networks, dropout, as approximate Bayesian inference, we perform a clear comparison proving the efficiency of Bayesian deep learning to maintain state-of-the-art performance compared to existing methods while mitigating the problem of uncertainty in deep learning.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.014 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.004 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.002 | 0.004 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".