We develop a new method to estimate failure probabilities in complex systems.
problem Estimating failure probabilities in safety-critical autonomous systems is challenging due to the rarity of failures and large state spaces.
method We propose an adaptive importance sampling algorithm that minimizes forward Kullback-Leibler divergence and uses Markov score ascent methods.
result Our method provides more accurate failure probability estimates than existing techniques.
Active Kriging Monte Carlo simulation method with conformal certification for failure probability estimation
problem Failure probability estimation in structural reliability analysis
method Active learning framework with conformal prediction
result Improved uncertainty quantification and reliability of failure probability estimates
Entropy-based GP adaptive design improves failure probability estimation.
problem Limited accuracy in failure probability estimation due to model evaluation costs.
method Entropy-based Gaussian process (GP) adaptive design combined with multifidelity importance sampling (MFIS).
result More accurate failure probability estimates and higher confidence.
Improves FI-PINNs by combining re-sampling and subset simulation for better failure probability estimation.
problem Estimating failure probability in physics-informed neural networks (PINNs).
method Adaptive sampling with re-sampling and subset simulation, using cosine-annealing for uniform to adaptive transition.
result Significant improvement in estimating failure probability and generating new training points in the failure region.
A new method estimates rare failure events in complex systems.
problem Estimating the probability of rare failure events in non-linear systems.
method Stochastic Spectral Embedding (SSE) combined with modifications for efficient rare event estimation.
result Rare failure probability decomposed into conditional probabilities for easier computation.
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…
Brain uses synaptic failure to sample from posterior distributions.
problem Bayesian inference in the brain's probabilistic computations.
method Adapting synaptic failure to sample posterior predictive distributions.
result Synaptic failure enables sampling of complete posterior predictive distributions.
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.
Method quantifies sensitivity of reliability analysis to uncertainty sources.
problem Computational expense in reliability analysis of complex models.
method Gaussian process surrogate model, active learning, sensitivity analysis.
result Reduces main source of error in estimating rare event probabilities.
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.
New method calculates sensitivity of system failure probability.
problem Difficulty in computing sensitivity of failure probability.
method Monte Carlo strategy using response gradient and kernel smoothing.
result Single Monte Carlo run for sensitivity estimates.
Paper introduces a novel method for estimating model confidence in deep neural classifiers.
problem Reliable confidence estimation for deep neural classifiers in safety-critical applications.
method Proposes a novel target criterion (true class probability) and learns it from data with an auxiliary model.
result The proposed method outperforms strong baselines in various tasks and network architectures.
Two BO methods improve reliability optimization for rare failures.
problem Maximizing reliability of designs subject to random perturbations.
method Bayesian optimization with Thompson sampling and knowledge gradient.
result Proposed methods outperform existing techniques in extreme failure probability scenarios.
While machine learning systems show high success rate in many complex tasks, research shows they can also fail in very unexpected situations. Rise of machine learning products in safety-critical industries cause an increase in attention in evaluating model robustness and estimating failure probability in machine learni…
Study constructs balanced datasets for seismic failure prediction.
problem Imbalanced datasets limit machine learning performance in seismic failure prediction.
method Framework with three steps: GMF identification, probability density estimation, and sample transformation.
result Framework improves machine learning performance in seismic failure mode prediction.
New framework tests AVs as a black box, prioritizing rare failure modes.
problem Lack of rigorous and scalable testing methods for AVs.
method Developed a simulation testing framework that learns to identify and rank failure scenarios via adaptive importance-sampling methods.
result First independent evaluation of a full-stack commercial AV system (Comma AI's OpenPilot).
The goal of this paper is to prove a result conjectured in Föllmer and Schachermayer [FS07], even in slightly more general form. Suppose that S is a continuous semimartingale and satisfies a large deviations estimate; this is a particular growth condition on the mean-variance tradeoff process of S. We show that S then …
A new machine learning method calculates failure probability efficiently and accurately.
problem Computing the probability of failure for complex systems.
method Penalized Profile Support Vector Machine with adaptive sampling and clustering.
result The method minimizes model evaluations while preserving decision boundary geometry.
In survival analysis, estimating the failure time distribution is an important and difficult task, since usually the data is subject to censoring. Specifically, in this paper we consider current status data, a type of data where all of the observations are censored. The format of the data is such that the failure time …
Adaptive PINNs improve accuracy by adding points where solutions are uncertain.
problem Inadequate sampling in PINNs leads to inaccurate solutions, especially near singularities.
method FI-PINNs use failure probability to dynamically add points, improving numerical accuracy.
result FI-PINNs achieve better accuracy through adaptive sampling, as proven by rigorous error bounds.
New method identifies subgroups in censored data.
problem Identifying meaningful patterns in heterogeneous populations.
method Combining inverse probability weighting, M-estimation, and concave pairwise fusion penalization.
result Robust approach for censored data under heterogeneous AFT models.
The aim of the present paper is to develop a strategy for solving reliability-based design optimization (RBDO) problems that remains applicable when the performance models are expensive to evaluate. Starting with the premise that simulation-based approaches are not affordable for such problems, and that the most-probab…
DG separates successes and failures by gating updates with advantage and surprisal.
problem Negative learning from surprising data in distributed reinforcement learning.
method DG gates each update with the product of advantage and surprisal, suppressing failures and preserving successes.
result DG outperforms other methods in various challenging reinforcement learning tasks.
We give the first efficient algorithm for learning the structure of an Ising model that tolerates independent failures; that is, each entry of the observed sample is missing with some unknown probability p. Our algorithm matches the essentially optimal runtime and sample complexity bounds of recent work for learning Is…
Hallucinations in models are mislinked estimates, not errors.
problem Hallucinations in generative models as failures to link estimates to plausible causes.
method Formalized hallucinations, showed even optimal estimators hallucinate, provided a general lower bound on hallucinate rate, reframed hallucination as structural misalignment, and experimentally supported theory.
result Hallucinations are structural misalignments between loss minimization and human-acceptable outputs, leading to estimation errors.
One of the key challenges in predictive maintenance is to predict the impending downtime of an equipment with a reasonable prediction horizon so that countermeasures can be put in place. Classically, this problem has been posed in two different ways which are typically solved independently: (1) Remaining useful life (R…
New method combines neural networks with Monte Carlo for complex system reliability.
problem Estimating small failure probabilities in complex systems.
method Subset Simulation with Hamiltonian Neural Networks.
result High acceptance rates and computational efficiency in low-probability regions.
Locally learned synaptic failure enables complete Bayesian inference.
problem Bayesian inference in neural networks.
method Biologically constrained neural network with synaptic failure and lateral inhibition.
result Synaptic failure allows sampling from both epistemic and aleatoric distributions.
Network analysis improves risk assessment for surety bonds.
problem Network effects in surety bonds increase risk assessment complexity.
method Modelled contractor network as directed graph, extended Friedkin-Johnsen model with stochastic process.
result Network effects increase average risk for surety organizations.
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…
This paper presents a method to efficiently estimate rare event probabilities using a combination of high and low-fidelity models.
problem Estimating the probability of failure for complex systems using high-fidelity models is expensive and inaccurate for rare events.
method The paper introduces a multi-fidelity surrogate modeling strategy using active learning and subset simulation to merge high and low-fidelity models.
result The method significantly reduces computational cost while maintaining high accuracy in estimating rare event probabilities.
We derive a closed form solution for an optimal control problem related to an interbank lending schemes subject to terminal probability constraints on the failure of banks which are interconnected through a financial network. The derived solution applies to a real banks network by obtaining a general solution when the …
ProEval efficiently estimates AI performance and discovers failures using pre-trained Gaussian Processes.
problem Resource-intensive evaluation of generative AI models.
method ProEval uses pre-trained Gaussian Processes and Bayesian quadrature to estimate performance and discover failures.
result ProEval requires significantly fewer samples to achieve accurate performance estimates and reveals more diverse failure cases.
Structural reliability methods aim at computing the probability of failure of systems with respect to some prescribed performance functions. In modern engineering such functions usually resort to running an expensive-to-evaluate computational model (e.g. a finite element model). In this respect simulation methods, whic…
Efficiently estimates rare events using multifidelity modeling.
problem Estimating rare events with computationally expensive models.
method Active learning with multifidelity modeling, adapting the number of high-fidelity simulations based on problem complexity and desired accuracy.
result Significantly reduced the number of high-fidelity model calls while maintaining accuracy.
K-means fails in high dimensions with noise and few samples.
problem Clustering in high-dimensional data with noise and limited samples.
method Simple Gaussian Mixture Model (GMM) analysis.
result Almost every partition becomes a fixed point of k-means in high dimensions.
OptIMIS method improves SRAM yield estimation efficiency and accuracy.
problem Efficient estimation of SRAM failure probability with shrinking technology nodes.
method Generalized norm minimization method, optimal manifold concept, onion sampling, neural coupling flow.
result OptIMIS method delivers up to 3.5x efficiency and 3x accuracy over state-of-the-art methods.
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.
Estimates reliability of nuclear fuel using advanced modeling techniques.
problem Determining the reliability of TRISO-coated particle fuel, which has small failure probabilities and expensive computational models.
method Coupled active learning, multifidelity modeling, and subset simulation.
result Multifidelity modeling strategies consistently reduce the number of high-fidelity model calls.
Credit networks represent a way of modeling trust between entities in a network. Nodes in the network print their own currency and trust each other for a certain amount of each other's currency. This allows the network to serve as a decentralized payment infrastructure---arbitrary payments can be routed through the net…
New sampling bounds improve uniform coverage verification in machine learning.
problem Conservative bounds in classical coverage analyses at small failure probabilities.
method Variance-based analysis of uniform random sampling on a d-dimensional unit hypercube. result Sample complexity bound with logarithmic dependence on failure probability.
The paper tackles safe exploration in RL by a conservative safety critic.
problem Safe exploration in reinforcement learning (RL) when partially trained policies are deployed.
method Learning a conservative safety estimate through a critic, provably bounding catastrophic failures.
result The approach provably converges to competitive task performance with significantly lower catastrophic failure rates.
When will a server fail catastrophically in an industrial datacenter? Is it possible to forecast these failures so preventive actions can be taken to increase the reliability of a datacenter? To answer these questions, we have studied what are probably the largest, publicly available datacenter traces, containing more …
New algorithm estimates eigenspace with faulty nodes, matching performance of existing methods.
problem Estimating eigenspace in distributed systems with node failures.
method Develops an eigenspace estimation algorithm for distributed environments with arbitrary node failures.
result Matches performance of existing non-robust estimator up to an additive error.
Method uses Seq2Seq learning to automatically generate recovery commands for ICT systems.
problem Manual decision-making for recovery commands is time-consuming and error-prone.
method Seq2Seq neural network model trained on past logs and commands.
result The model can estimate accurate recovery commands from new failures.
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.
Bayesian inference models failure distributions in autonomous systems.
problem Estimating the distribution of failures in complex systems.
method Bayesian inference using system dynamics rollouts and gradient computation.
result Improves sample efficiency and parameter space coverage in autonomous systems.
We propose a method to optimise the parameters of a policy which will be used to safely perform a given task in a data-efficient manner. We train a Gaussian process model to capture the system dynamics, based on the PILCO framework. Our model has useful analytic properties, which allow closed form computation of error …