Large language models (LLMs) are prone to hallucinations, i.e., statements unsupported by the input or training data, hindering reliable deployment. In parallel, numerous uncertainty estimation (UE) methods have been proposed to quantify model confidence and are often implicitly treated as proxies for model failure. However, the relationship between uncertainty and hallucinations remains insufficiently characterized. We present a systematic empirical study of the association between uncertainty estimators and hallucinations in LLMs. Rather than assuming this association, we evaluate directly when and to what extent it holds. We consider a diverse set of uncertainty estimators, including information-theoretic, sampling-based, and reflexive estimators, and examine their behavior across hallucination settings. Our experiments cover both intrinsic hallucinations (violations of input faithfulness) and extrinsic hallucinations (unsupported claims relative to training data), using four complementary benchmarks, including RAGTruth and HalluLens. We find that the association is highly variable and often weak, depending on the hallucination type and the LLM under evaluation. These results challenge the use of uncertainty as a direct signal of hallucination and clarify when it provides actionable information.

Contemporary vision benchmarks predominantly consider tasks on which humans can achieve near-perfect performance. However, humans are frequently presented with visual data that they cannot classify with 100% certainty, and models trained on standard vision benchmarks achieve low performance when evaluated on this data. To address this issue, we introduce a procedure for creating datasets of ambiguous images and use it to produce SQUID-E ("Squidy"), a collection of noisy images extracted from videos. All images are annotated with ground truth values and a test set is annotated with human uncertainty judgments. We use this dataset to characterize human uncertainty in vision tasks and evaluate existing visual event classification models. Experimental results suggest that existing vision models are not sufficiently equipped to provide meaningful outputs for ambiguous images and that datasets of this nature can be used to assess and improve such models through model training and direct evaluation of model calibration. These findings motivate large-scale ambiguous dataset creation and further research focusing on noisy visual data.1

Current uncertainty quantification is memory and compute expensive, which hinders practical uptake. To counter, we develop Sketched Lanczos Uncertainty (SLU): an architecture-agnostic uncertainty score that can be applied to pre-trained neural networks with minimal overhead. Importantly, the memory use of SLU only grows logarithmically with the number of model parameters. We combine Lanczos' algorithm with dimensionality reduction techniques to compute a sketch of the leading eigenvectors of a matrix. Applying this novel algorithm to the Fisher information matrix yields a cheap and reliable uncertainty score. Empirically, SLU yields well-calibrated uncertainties, reliably detects out-of-distribution examples, and consistently outperforms existing methods in the low-memory regime.

Uncertainty estimation aims to evaluate the confidence of a trained deep neural network.

Large language models(LLMs) are increasingly expanding their real-world applications across domains, e.g., question answering, autonomous driving, and automatic software development. Despite this achievement, LLMs, as data-driven systems, often make incorrect predictions, which can lead to potential losses in safety-critical scenarios. To address this issue and measure the confidence of model outputs, multiple uncertainty quantification(UQ) criteria have been proposed. However, even though important, there are limited tools to integrate these methods, hindering the practical usage of UQ methods and future research in this domain. To bridge this gap, in this paper, we introduce UncertaintyZoo, a unified toolkit that integrates 29 uncertainty quantification methods, covering five major categories under a standardized interface. Using UncertaintyZoo, we evaluate the usefulness of existing uncertainty quantification methods under the code vulnerability detection task on CodeBERT and ChatGLM3 models. The results demonstrate that UncertaintyZoo effectively reveals prediction uncertainty. The tool with a demonstration video is available on the project site https://github.com/Paddingbuta/UncertaintyZoo.

The advance of Large Language Models (LLMs) has greatly stimulated research interest in developing multi-modal LLM (MLLM)-based visual anomaly detection (VAD) algorithms that can be deployed in complex environments. The challenge is that in these complex environments, the anomalies are sometimes highly contextual and also ambiguous, and thereby, uncertainty quantification (UQ) is a crucial capacity for an MLLM-based VAD system to succeed. In this paper, we introduce our UQ-supported MLLM-based VAD framework called ALARM. ALARM integrates UQ with quality-assurance techniques like reasoning chain, self-reflection, and MLLM ensemble for robust and accurate performance and is designed based on a rigorous probabilistic inference pipeline and computational process. Extensive empirical evaluations are conducted using the real-world smart-home benchmark data and wound image classification data, which shows ALARM's superior performance and its generic applicability across different domains for reliable decision-making.

Ensuring the safety of real-world systems is challenging, especially when they rely on learned perception modules to infer the system state from high-dimensional sensor data. These perception modules are vulnerable to epistemic uncertainty, often failing when encountering out-of-distribution (OoD) measurements not seen during training. To address this gap, we introduce ATOM-CBF (Adaptive-To-OoD-Measurement Control Barrier Function), a novel safe control framework that explicitly computes and adapts to the epistemic uncertainty from OoD measurements, without the need for ground-truth labels or information on distribution shifts. Our approach features two key components: (1) an OoD-aware adaptive perception error margin and (2) a safety filter that integrates this adaptive error margin, enabling the filter to adjust its conservatism in real-time. We provide empirical validation in simulations, demonstrating that ATOM-CBF maintains safety for an F1Tenth vehicle with LiDAR scans and a quadruped robot with RGB images.

Obtaining high-quality labels for large datasets is expensive, requiring massive annotations from human experts. While AI models offer a cost-effective alternative by predicting labels, their label quality is compromised by the unavoidable labeling errors. Existing methods mitigate this issue through selective labeling, where AI labels a subset and human labels the remainder. However, these methods lack theoretical guarantees on the quality of AI-assigned labels, often resulting in unacceptably high labeling error within the AI-labeled subset. To address this, we introduce \textbf{Conformal Labeling}, a novel method to identify instances where AI predictions can be provably trusted. This is achieved by controlling the false discovery rate (FDR), the proportion of incorrect labels within the selected subset. In particular, we construct a conformal $p$-value for each test instance by comparing AI models' predicted confidence to those of calibration instances mislabeled by AI models. Then, we select test instances whose $p$-values are below a data-dependent threshold, certifying AI models' predictions as trustworthy. We provide theoretical guarantees that Conformal Labeling controls the FDR below the nominal level, ensuring that a predefined fraction of AI-assigned labels is correct on average. Extensive experiments demonstrate that our method achieves tight FDR control with high power across various tasks, including image and text labeling, and LLM QA.