Dropout Robustness and Cognitive Profiling of Transformer Models via Stochastic Inference

• URL: http://arxiv.org/abs/2603.17811v1
• Date: Wed, 18 Mar 2026 15:04:26 GMT
• Title: Dropout Robustness and Cognitive Profiling of Transformer Models via Stochastic Inference
• Abstract summary: Transformer-based language models are widely deployed for reasoning, yet their behavior under inference-timeity remains underexplored.<n>This work analyzes dropout-induced variability across 19 transformer models using MC Dropout with 100 forward passes per sample.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Transformer-based language models are widely deployed for reasoning, yet their behavior under inference-time stochasticity remains underexplored. While dropout is common during training, its inference-time effects via Monte Carlo sampling lack systematic evaluation across architectures, limiting understanding of model reliability in uncertainty-aware applications. This work analyzes dropout-induced variability across 19 transformer models using MC Dropout with 100 stochastic forward passes per sample. Dropout robustness is defined as maintaining high accuracy and stable predictions under stochastic inference, measured by standard deviation of per-run accuracies. A cognitive decomposition framework disentangles performance into memory and reasoning components. Experiments span five dropout configurations yielding 95 unique evaluations on 1,000 samples. Results reveal substantial architectural variation. Smaller models demonstrate perfect prediction stability while medium-sized models exhibit notable volatility. Mid-sized models achieve the best overall performance; larger models excel at memory tasks. Critically, 53% of models suffer severe accuracy degradation under baseline MC Dropout, with task-specialized models losing up to 24 percentage points, indicating unsuitability for uncertainty quantification in these architectures. Asymmetric effects emerge: high dropout reduces memory accuracy by 27 percentage points while reasoning degrades only 1 point, suggesting memory tasks rely on stable representations that dropout disrupts. 84% of models demonstrate memory-biased performance. This provides the first comprehensive MC Dropout benchmark for transformers, revealing dropout robustness is architecture-dependent and uncorrelated with scale. The cognitive profiling framework offers actionable guidance for model selection in uncertainty-aware applications.

Related papers

• Accuracy and Robustness of Model Cascades Under Data Perturbations [2.1920579994942164]
  Prediction cascades significantly reduce energy consumption of Artificial Intelligence (AI) models while maintaining high predictive performance.<n>We study confidence-based cascade frameworks for image classification and investigate how such degradations affect their confidence-based deferral behavior.
  arXiv  Detail & Related papers  (2026-08-18T12:34:09Z)
• Joint Model and Data Sparsification via the Marginal Likelihood [53.29070892356214]
  We propose jointly learning individual feature and sample relevancies, enabling simultaneous model and data sparsification.<n>This symmetric pruning of model and data offers a natural extension that preserves conjugacy.<n> Empirical results across diverse regression tasks affirm that a joint ARD approach consistently yields both sparse and robust prediction models.
  arXiv  Detail & Related papers  (2026-05-28T13:26:53Z)
• Fortress: A Case Study in Stabilizing Search Recommendations via Temporal Data Augmentation and Feature Pruning [2.6494136490021387]
  Fortress identifies and prunes features that contribute to inconsistent prediction scores over time.<n>We validate Fortress on a query-to-app relevance model in a large-scale app marketplace.
  arXiv  Detail & Related papers  (2026-05-14T18:13:05Z)
• Evaluating Uplift Modeling under Structural Biases: Insights into Metric Stability and Model Robustness [8.135022024189306]
  In personalized marketing, uplift models estimate incremental effects by modeling how customer behavior changes under alternative treatments.<n>Real-world data often exhibit biases, such as selection bias, spillover effects, and unobserved confounding.
  arXiv  Detail & Related papers  (2026-03-21T11:54:25Z)
• Attribution-Guided Model Rectification of Unreliable Neural Network Behaviors [60.06461883533697]
  We leverage rank-one model editing to establish an attribution-guided model rectification framework.<n>We first distinguish our rectification setting from existing model editing, yielding a formulation that corrects unreliable behavior.<n>To target the primary source of misbehavior, we introduce an attribution-guided layer localization method.
  arXiv  Detail & Related papers  (2026-03-08T01:06:07Z)
• DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting [59.868414584142336]
  DropoutTS is a model-agnostic plugin that shifts the paradigm from "what" to "how much" to learn.<n>It maps noise to adaptive dropout rates - selectively suppressing spurious fluctuations while preserving fine-grained fidelity.
  arXiv  Detail & Related papers  (2026-01-29T13:49:20Z)
• The Drill-Down and Fabricate Test (DDFT): A Protocol for Measuring Epistemic Robustness in Language Models [0.0]
  Current language model evaluations measure what models know under ideal conditions but not how robustly they know it under realistic stress.<n>We introduce the Drill-Down Fabricate Test (DDFT), a protocol that measures robustness.<n>We find flagship models exhibit brittleness despite their scale, while smaller models can achieve robust performance.
  arXiv  Detail & Related papers  (2025-12-29T20:29:09Z)
• SG-OIF: A Stability-Guided Online Influence Framework for Reliable Vision Data [6.4391040754741296]
  In this paper, we introduce a Stability-Guided Online Influence Framework (SG-OIF) for Approximating training-point influence on test predictions.<n>We show that SG-OIF achieves 91.1% accuracy in the top 1% prediction samples on the CIFAR-10, and 99.8% AUPR score on MNIST.
  arXiv  Detail & Related papers  (2025-11-21T19:58:54Z)
• Uncertainty-aware Human Mobility Modeling and Anomaly Detection [24.22648449430148]
  We formulate anomaly detection in human behavior modeling raw GPS data as sequence stay-point events.<n>We equip our proposed model USTAD with aleatoric uncertainty estimation.<n>Experiments show that USTAD improves anomaly detection AUCROC by 3%-15% over baselines in industry-scale data.
  arXiv  Detail & Related papers  (2024-10-02T06:57:08Z)
• Selective Learning: Towards Robust Calibration with Dynamic Regularization [79.92633587914659]
  Miscalibration in deep learning refers to there is a discrepancy between the predicted confidence and performance. We introduce Dynamic Regularization (DReg) which aims to learn what should be learned during training thereby circumventing the confidence adjusting trade-off.
  arXiv  Detail & Related papers  (2024-02-13T11:25:20Z)
• Towards a performance characteristic curve for model evaluation: an application in information diffusion prediction [2.8686437689115354]
  We propose a metric based on information entropy to quantify the randomness in diffusion data.<n>We then identify a scaling pattern between the randomness and the prediction accuracy of the model.<n>The validity of the curve is tested by three prediction models in the same family.
  arXiv  Detail & Related papers  (2023-09-18T07:32:57Z)
• Learning Sample Difficulty from Pre-trained Models for Reliable Prediction [55.77136037458667]
  We propose to utilize large-scale pre-trained models to guide downstream model training with sample difficulty-aware entropy regularization. We simultaneously improve accuracy and uncertainty calibration across challenging benchmarks.
  arXiv  Detail & Related papers  (2023-04-20T07:29:23Z)
• Robustness and Accuracy Could Be Reconcilable by (Proper) Definition [109.62614226793833]
  The trade-off between robustness and accuracy has been widely studied in the adversarial literature. We find that it may stem from the improperly defined robust error, which imposes an inductive bias of local invariance. By definition, SCORE facilitates the reconciliation between robustness and accuracy, while still handling the worst-case uncertainty.
  arXiv  Detail & Related papers  (2022-02-21T10:36:09Z)
• Contextual Dropout: An Efficient Sample-Dependent Dropout Module [60.63525456640462]
  Dropout has been demonstrated as a simple and effective module to regularize the training process of deep neural networks. We propose contextual dropout with an efficient structural design as a simple and scalable sample-dependent dropout module. Our experimental results show that the proposed method outperforms baseline methods in terms of both accuracy and quality of uncertainty estimation.
  arXiv  Detail & Related papers  (2021-03-06T19:30:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.