Risk Advisor predicts and mitigates ML deployment failures.
problem Predicting and mitigating test-time failure risks of ML systems.
method Post-hoc meta-learner for estimating failure risks and uncertainties.
result Reliably predicts deployment-time failure risks across various ML models.
PAGER detects failures in deep regression models using a new framework.
problem Detecting failures in deep regression models.
method PAGER uses a combination of epistemic uncertainty and manifold non-conformity scores.
result PAGER accurately characterizes and detects failures in deep regressors.
Underspecified ML models can behave unpredictably in real-world use.
problem ML models can fail in real-world deployment due to ambiguous predictors.
method Identified underspecification as the cause, showing it affects various ML domains.
result Underspecified models can behave differently in deployment domains.
CalNF models rare failures with limited data, improving safety in autonomous systems.
problem Challenges in modeling and debugging rare safety-critical failures due to limited data.
method CalNF, a self-regularized framework for posterior learning from limited data.
result Achieves state-of-the-art performance on data-limited failure modeling and inverse problems.
This paper addresses the problem of evaluating learning systems in safety critical domains such as autonomous driving, where failures can have catastrophic consequences. We focus on two problems: searching for scenarios when learned agents fail and assessing their probability of failure. The standard method for agent e…
Deep neural networks predict CVCM track circuit failures early.
problem Subtle anomalies in CVCM track circuits lead to failures, causing disruptions.
method Deep neural networks classify anomalies before they escalate.
result Deep neural networks achieve 99.31% overall accuracy in detecting CVCM failures.
Assessing reliably the confidence of a deep neural network and predicting its failures is of primary importance for the practical deployment of these models. In this paper, we propose a new target criterion for model confidence, corresponding to the True Class Probability (TCP). We show how using the TCP is more suited…
Diverging Flows detects extrapolations in flow models, ensuring reliable predictions.
problem Flow models extrapolate into invalid data, leading to silent failures.
method Structurally enforce inefficient transport for off-manifold inputs.
result Effective detection of extrapolations without compromising predictive fidelity or inference latency.
The traditional paradigm for developing machine prognostics usually relies on generalization from data acquired in experiments under controlled conditions prior to deployment of the equipment. Detecting or predicting failures and estimating machine health in this way assumes that future field data will have a very simi…
New framework assesses extreme errors in machine learning models.
problem Current validation methods fail to quantify extreme errors in high-stakes domains.
method Uses Extreme Value Theory (EVT) to estimate worst-case failures.
result Establishes EVT as a fundamental tool for assessing model reliability.
Suitability filter detects model performance degradation in real-world deployment.
problem Ensuring model reliability in safety-critical domains without access to ground truth labels.
method Uses suitability signals to evaluate classifier performance on unlabeled user data.
result The suitability filter reliably detects performance deviations due to covariate shift.
The study diagnoses fairness issues in healthcare models under distribution shifts.
problem Understanding and diagnosing fairness changes in machine learning models under distribution shifts in healthcare.
method Causal framing and conditional independence tests to characterize distribution shifts.
result Knowledge of distribution shifts helps diagnose fairness transfer failures, including complex cases.
Develops confidence bounds for off-policy evaluation in contextual bandits.
problem Evaluating policies that were not used to collect data.
method Martingale analysis for non-asymptotic, non-parametric, and valid confidence sequences.
result Empirically tight bounds on failure probability and width.
Proposes integrating global and local entropy for more reliable LLMs.
problem Uncertainty in large language models (LLMs) leads to unreliable predictions.
method Measures global uncertainty from hidden-state matrices and local uncertainty from tokens, combining them via a multiplicative gate.
result Global-Local Uncertainty (GLU) outperforms unsupervised baselines across multiple models and benchmarks.
Joint training improves model accuracy by selectively using privileged information.
problem Two-stage training can lead to model failure with noisy privileged information.
method Joint training of two models to use privileged information selectively.
result Joint training outperforms two-stage baselines on synthetic and real-world tasks.
Bayesian optimization improves classifier selection for acute infection and mortality.
problem Improving accuracy of acute infection and mortality prediction.
method Comparison of hyperparameter optimization methods (grid search, random sampling, Bayesian optimization).
result Bayesian optimization outperforms grid search or random sampling for in-hospital mortality classifiers.
New framework assesses LLM security risks in BFSI.
problem Lack of domain-specific security evaluation for LLMs in BFSI.
method Risk-aware evaluation framework combining taxonomy, automated red-teaming, and ensemble judging.
result Higher decoding stochasticity and adaptive interaction lead to more severe disclosures.
CRC improves multivariate forecasting accuracy without risking performance degradation.
problem Systematic errors and lack of guarantees in multivariate forecasters.
method CRC uses a causality-inspired encoder and hybrid corrector with a safety mechanism.
result CRC consistently improves accuracy and ensures high non-degradation rates.
Automatic methods for generating state-of-the-art neural network architectures without human experts have generated significant attention recently. This is because of the potential to remove human experts from the design loop which can reduce costs and decrease time to model deployment. Neural architecture search (NAS)…
A geometric theory explains loss functions for robust representation learning.
problem Treats robustness, domain adaptation, and sensor drift as separate literatures.
method Estimates covariance Sigma_task and uses it to pin Jacobian penalties.
result Proves optimality and necessity of range coverage for penalty matrices.
PRoA assesses deep learning robustness against practical functional perturbations.
problem Inadequate practical robustness verification methods for deep learning systems.
method Probabilistic robustness assessment based on adaptive concentration.
result Statistical guarantees on probabilistic robustness against functional perturbations.
Mixture models are a fundamental tool in applied statistics and machine learning for treating data taken from multiple subpopulations. The current practice for estimating the parameters of such models relies on local search heuristics (e.g., the EM algorithm) which are prone to failure, and existing consistent methods …
Deployment-complete benchmarking assesses if evidence leads to consistent deployment actions.
problem Lack of clear evidence leading to consistent deployment actions.
method Introduces deployment-complete benchmarking to test if benchmark evidence determines deployment actions.
result Benchmark evidence must be complete for a claim to lead to a consistent deployment action.
A new method for releasing AI workflows to avoid premature incorrect results.
problem Statistical challenges in releasing AI workflows with adaptive scoring.
method Wrapper that calibrates and accumulates evidence from high-scoring failures.
result Reduces premature incorrect release while still releasing on moderate evidence.
Paper establishes lower bounds and optimal algorithms for deployment-efficient RL.
problem Deployment efficiency in reinforcement learning.
method Optimization with constraints, lower bounds, algorithms.
result Established optimal algorithms for deployment-efficient RL.
Traditional control methods are inadequate in many deployment settings involving control of Cyber-Physical Systems (CPS). In such settings, CPS controllers must operate and respond to unpredictable interactions, conditions, or failure modes. Dealing with such unpredictability requires the use of executive and cognitive…
DUPLE tackles cross-deployment recognition in fiber-optic perimeter security with meta-learning.
problem Cross-deployment recognition challenges in fiber-optic perimeter security due to label scarcity and distribution shifts.
method DUPLE employs statistically guided meta-learning to enhance recognition robustness across unseen deployments.
result DUPLE consistently outperforms traditional and meta-learning baselines in cross-deployment DFOS benchmarks.
Personalized models using group attributes reduce performance, study finds.
problem Reducing performance of models using group attributes like race or gender.
method Formal conditions and collective preference guarantees to ensure fair use.
result Models personalized with group attributes reduce performance at a group level.
Machine learning (ML), especially deep learning is made possible by the availability of big data, enormous compute power and, often overlooked, development tools or frameworks. As the algorithms become mature and efficient, more and more ML inference is moving out of datacenters/cloud and deployed on edge devices. This…
Economics tool predicts failure times in reliability systems.
problem Predicting optimal failure times in weighted k-out-of-n reliability systems with heterogeneous component failure.
method Using rational expectations to analyze and predict failure times in reliability systems with heterogeneous component failure.
result Different measures are optimal for predicting system failure depending on component failure distributions.
Unified model predicts multi-mode failure with multi-sensor data.
problem Independent failure mode and RUL prediction ignores inherent relationship.
method Hierarchical Bayesian framework with Cox model, Gaussian process, and multinomial distributions.
result Robust uncertainty quantification and accurate prediction of multi-mode failure.
Framework for responsible LLM deployment with human involvement and decentralized technologies.
problem Challenges in deploying LLMs for high-stakes decisions, including data security and accountability.
method Interactive human involvement through multiple iterations, decentralized technologies, and automated auditing.
result Enhanced security and accountability in LLM deployment for financial decisions.
The study examines how language models learn to represent the world, identifying conditions for ecological veridicality.
problem Understanding when language models learn to represent the world accurately and how this learning process can fail.
method Analyzes the Bayes-optimal next-token cross-entropy decomposition and the role of training ecology in shaping model representations.
result The minimum-complexity zero-excess solution is the quotient partition by training equivalence, and this solution is not preserved in in-context learning or per-task adaptation.
Prognostics and Health Management (PHM) is an emerging engineering discipline which is concerned with the analysis and prediction of equipment health and performance. One of the key challenges in PHM is to accurately predict impending failures in the equipment. In recent years, solutions for failure prediction have evo…
This work improves safety validation of autonomous vehicles by finding interpretable failures.
problem Finding interpretable failures of autonomous systems in simulation.
method Signal temporal logic expressions optimized for high likelihood and human interpretability.
result Our methodology finds more interpretable failures with higher likelihood compared to baseline approaches.
Framework predicts remaining useful life of DSH subsystems under unknown failure modes.
problem Predicting remaining useful life of DSH subsystems with unknown failure modes.
method Unsupervised framework using mixture of Gaussian regressions and Expectation-Maximization algorithm.
result Improved prediction accuracy and interpretability of RUL.
This thesis tackles data imperfections in ML, proposing methods to prevent discrimination and spurious feature learning.
problem Data imperfections in ML training data, leading to deployment failures.
method Fair representation learning, spurious feature detection, and data augmentation for reinforcement learning.
result Methods to make ML models more robust to data imperfections.
The paper calculates the likelihood of a financial market failure involving multiple major banks.
problem Estimating the probability of a market failure involving multiple globally important banks.
method Multivariate Cox process across G-SIBs, deriving various theorems on market failure probabilities.
result The probability of a market failure increases with the number of G-SIBs and is inevitable if there are too many.
Conformal prediction fails under severe feature turnover in COVID-19 supply chain tasks.
problem Dealing with distribution shift in conformal prediction models.
method Using COVID-19 as a natural experiment across 8 supply chain tasks, analyzing SHAP explanations.
result Coverage drops vary widely (0% to 86.7%) and correlate with single-feature dependence.
Paper proposes a method to predict disk failures using multi-layer domain adaptive learning.
problem Traditional machine learning models struggle to predict disk failures due to limited data.
method Multi-layer domain adaptive learning with source and target domains.
result The proposed method improves failure prediction accuracy on disk data with few failure samples.
CSP improves time-series forecasting without training, outperforming DeepNPTS in speed and accuracy.
problem Improving probabilistic time-series forecasting without training.
method Mixing empirical and residual draws around a seasonal naive forecast.
result CSP significantly outperforms DeepNPTS on CRPS, normalized mean quantile loss, and coverage metrics.
We aim to predict and explain service failures in supply-chain networks, more precisely among last-mile pickup and delivery services to customers. We analyze a dataset of 500,000 services using (1) supervised classification with Random Forests, and (2) Association Rules. Our classifier reaches an average sensitivity of…
GE finds failures in autonomous systems without domain heuristics.
problem Finding failures in autonomous systems without domain-specific heuristics.
method Adaptive stress testing using go-explore (GE) algorithm.
result GE finds failures in scenarios other RL techniques cannot solve.
A method for quickly determining deployment schedules that meet a given fuel cycle demand is presented here. This algorithm is fast enough to perform in situ within low-fidelity fuel cycle simulators. It uses Gaussian process regression models to predict the production curve as a function of time and the number of depl…
The paper develops methods to reduce deployment risk under dynamic covariate shifts.
problem Reduction of deployment risk under dynamic covariate shifts.
method Time-domain Poincare inequality and Jacobian-velocity theorem to identify and control directional tangent energy.
result Drift-aligned tangent regularization (DTR) reduces risk volatility and directional gain in low-rank drift regimes.
New multi-spiral approach improves packaging of thick membranes.
problem Deploying thick membranes on curved surfaces efficiently.
method Multi-spiral folding approach with prismatic folding lines.
result Improved deployment performance of thick membranes on curved surfaces.
Method distinguishes between failures and domain shifts in industrial data streams.
problem Confusing domain shifts with failures in industrial data.
method Modified Page-Hinkley changepoint detector and supervised domain-adaptation-based anomaly detection.
result Allows differentiation between failures and domain shifts.
RL research overhypes potential but lacks deployable solutions.
problem Current RL research overhypes potential and lacks deployable solutions.
method Identifies and critiques current RL research practices.
result Current RL research direction is unlikely to lead to practical, economically viable solutions.