Author name cluster

Eric Nalisnick

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

13 papers

1 author row

TMLR Journal 2026 Journal Article

Are vision language models robust to classic uncertainty challenges?

Xi Wang
Eric Nalisnick

Robustness against uncertain and ambiguous inputs is a critical challenge for deep learning models. While recent advancements in large scale vision language models (VLMs, e.g. GPT-4o) might suggest that increasing model and training dataset size would mitigate this issue, our empirical evaluation shows a more complicated picture. In this work, we sanity check whether modern VLMs pass the two most ``classic'' uncertainty quantification challenges: Anomaly detection and classification under inherently ambiguous conditions, we find that newer and larger VLMs indeed exhibit improved robustness compared to earlier models, but still suffer from a tendency to strictly follow instructions, often causing them to hallucinate confident responses even when faced with unclear or anomalous inputs. Remarkably, for natural images such as ImageNet, this limitation can be overcome without pipeline modifications: simply prompting models to abstain from uncertain predictions enables significant reliability gains, achieving near-perfect robustness in several settings. However, for domain-specific tasks such as galaxy morphology classification, a lack of specialized knowledge prevents reliable uncertainty estimation. Finally, we propose a simple mechanism based on caption diversity to reveal a model’s internal uncertainty, enabling practitioners to predict when models will successfully abstain without relying on labeled data.

PDF Details

TMLR Journal 2026 Journal Article

Learning to Defer with an Uncertain Rejector via Conformal Prediction

Yizirui Fang
Eric Nalisnick

Learning to defer (L2D) aims to optimize human-AI collaboration by allocating prediction tasks to either a machine learning model or a human expert, depending on which is most likely to be correct. This allocation decision is governed by a rejector: a meta-model that routes inputs based on estimated success probabilities. In practice, a poorly fit or otherwise misspecified rejector can jeopardize the entire L2D workflow due to its crucial role in allocating prediction tasks. In this work, we perform uncertainty quantification for the rejector. We use conformal prediction to allow the rejector to output prediction sets or intervals instead of just the binary outcome of ‘defer’ or not. On tasks ranging from image to hate speech classification, we demonstrate that the uncertainty in the rejector translates to safer decisions via two forms of selective prediction.

PDF Details

JMLR Journal 2025 Journal Article

Lightning UQ Box: Uncertainty Quantification for Neural Networks

Nils Lehmann
Nina Maria Gottschling
Jakob Gawlikowski
Adam J. Stewart
Stefan Depeweg
Eric Nalisnick

Although neural networks have shown impressive results in a multitude of application domains, the "black box" nature of deep learning and lack of confidence estimates have led to scepticism, especially in domains like medicine and physics where such estimates are critical. Research on uncertainty quantification (UQ) has helped elucidate the reliability of these models, but existing implementations of these UQ methods are sparse and difficult to reuse. To this end, we introduce Lightning UQ Box, a PyTorch-based Python library for deep learning-based UQ methods powered by PyTorch Lightning. Lightning UQ Box supports classification, regression, semantic segmentation, and pixelwise regression applications, and UQ methods from a variety of theoretical motivations. With this library, we provide an entry point for practitioners new to UQ, as well as easy-to-use components and tools for scalable deep learning applications. [abs] [ pdf ][ bib ] [ code ] &copy JMLR 2025. ( edit, beta )

PDF Details

NeurIPS Conference 2025 Conference Paper

Monitoring Risks in Test-Time Adaptation

Mona Schirmer
Metod Jazbec
Christian Andersson Naesseth
Eric Nalisnick

Encountering shifted data at test time is a ubiquitous challenge when deploying predictive machine learning models. Test-time adaptation (TTA) methods aim to address this issue by continuously adapting a deployed model using only unlabeled test data. While TTA can help extend the model's deployment lifespan, there are scenarios where, despite adaptation, the drop in the model's performance remains significant enough to warrant taking the model offline and retraining. To detect such failure cases, we propose pairing TTA with risk monitoring frameworks that track predictive performance and raise alerts when predefined performance criteria are violated. Specifically, we extend existing monitoring tools based on sequential testing with confidence sequences to accommodate scenarios where the model is updated at test time and no test labels are available to estimate the performance metrics of interest. Our extensions unlock the application of rigorous statistical risk monitoring in TTA and we demonstrate applicability of our proposed TTA monitoring framework across a representative set of TTA methods, datasets and distribution shift types.

PDF Details

TMLR Journal 2025 Journal Article

On the Challenges and Opportunities in Generative AI

Laura Manduchi
Clara Meister
Kushagra Pandey
Robert Bamler
Ryan Cotterell
Sina Däubener
Sophie Fellenz
Asja Fischer

The field of deep generative modeling has grown rapidly in the last few years. With the availability of massive amounts of training data coupled with advances in scalable unsupervised learning paradigms, recent large-scale generative models show tremendous promise in synthesizing high-resolution images and text, as well as structured data such as videos and molecules. However, we argue that current large-scale generative AI models exhibit several fundamental shortcomings that hinder their widespread adoption across domains. In this work, our objective is to identify these issues and highlight key unresolved challenges in modern generative AI paradigms that should be addressed to further enhance their capabilities, versatility, and reliability. By identifying these challenges, we aim to provide researchers with insights for exploring fruitful research directions, thus fostering the development of more robust and accessible generative AI solutions.

PDF Details

TMLR Journal 2025 Journal Article

Scalable Generative Modeling of Weighted Graphs

Richard Williams
Eric Nalisnick
Andrew Holbrook

Weighted graphs are ubiquitous throughout biology, chemistry, and the social sciences, motivating the development of generative models for abstract weighted graph data using deep neural networks. However, most current deep generative models are designed for unweighted graphs and cannot be easily extended to weighted topologies. Among those that do incorporate edge weights, few consider a joint distribution with the topology of the graph. Furthermore, learning a distribution over weighted graphs must account for complex nonlocal dependencies between both the edges of the graph and corresponding weights of each edge. We develop an autoregressive model BiGG-E, a nontrivial extension of the BiGG model, that learns a joint distribution over weighted graphs while exploiting sparsity to generate a weighted graph with $n$ nodes and $m$ edges in $O((n + m)\log n)$ time. Simulation studies and experiments on a variety of benchmark datasets demonstrate that BiGG-E best captures distributions over weighted graphs while remaining scalable and computationally efficient.

PDF Details

TMLR Journal 2025 Journal Article

Temporal Test-Time Adaptation with State-Space Models

Mona Schirmer
Dan Zhang
Eric Nalisnick

Distribution shifts between training and test data are inevitable over the lifecycle of a deployed model, leading to performance decay. Adapting a model on test samples can help mitigate this drop in performance. However, most test-time adaptation methods have focused on synthetic corruption shifts, leaving a variety of distribution shifts underexplored. In this paper, we focus on distribution shifts that evolve gradually over time, which are common in the wild but challenging for existing methods, as we show. To address this, we propose STAD, a Bayesian filtering method that adapts a deployed model to temporal distribution shifts by learning the time-varying dynamics in the last set of hidden features. Without requiring labels, our model infers time-evolving class prototypes that act as a dynamic classification head. Through experiments on real-world temporal distribution shifts, we show that our method excels in handling small batch sizes and label shift.

PDF Details

NeurIPS Conference 2024 Conference Paper

A Generative Model of Symmetry Transformations

James U. Allingham
Bruno K. Mlodozeniec
Shreyas Padhy
Javier Antorán
David Krueger
Richard E. Turner
Eric Nalisnick
José M. Hernández-Lobato

Correctly capturing the symmetry transformations of data can lead to efficient models with strong generalization capabilities, though methods incorporating symmetries often require prior knowledge. While recent advancements have been made in learning those symmetries directly from the dataset, most of this work has focused on the discriminative setting. In this paper, we take inspiration from group theoretic ideas to construct a generative model that explicitly aims to capture the data's approximate symmetries. This results in a model that, given a prespecified broad set of possible symmetries, learns to what extent, if at all, those symmetries are actually present. Our model can be seen as a generative process for data augmentation. We provide a simple algorithm for learning our generative model and empirically demonstrate its ability to capture symmetries under affine and color transformations, in an interpretable way. Combining our symmetry model with standard generative models results in higher marginal test-log-likelihoods and improved data efficiency.

PDF Details DOI

NeurIPS Conference 2024 Conference Paper

Fast yet Safe: Early-Exiting with Risk Control

Metod Jazbec
Alexander Timans
Tin H. Veljković
Kaspar Sakmann
Dan Zhang
Christian A. Naesseth
Eric Nalisnick

Scaling machine learning models significantly improves their performance. However, such gains come at the cost of inference being slow and resource-intensive. Early-exit neural networks (EENNs) offer a promising solution: they accelerate inference by allowing intermediate layers to exit and produce a prediction early. Yet a fundamental issue with EENNs is how to determine when to exit without severely degrading performance. In other words, when is it 'safe' for an EENN to go 'fast'? To address this issue, we investigate how to adapt frameworks of risk control to EENNs. Risk control offers a distribution-free, post-hoc solution that tunes the EENN's exiting mechanism so that exits only occur when the output is of sufficient quality. We empirically validate our insights on a range of vision and language tasks, demonstrating that risk control can produce substantial computational savings, all the while preserving user-specified performance goals.

PDF Details DOI

TMLR Journal 2024 Journal Article

Practical Synthesis of Mixed-Tailed Data with Normalizing Flows

Saba Amiri
Eric Nalisnick
Adam Belloum
Sander Klous
Leon Gommans

Capturing the correct tail behavior is difficult, yet essential for a faithful generative model. In this work, we provide an improved framework for training flows-based models with robust capabilities to capture the tail behavior of mixed-tail data. We propose a combination of a tail-flexible base distribution and a robust training algorithm to enable the flow to model heterogeneous tail behavior in the target distribution. We support our claim with extensive experiments on synthetic and real world data.

PDF Details

NeurIPS Conference 2023 Conference Paper

Towards Anytime Classification in Early-Exit Architectures by Enforcing Conditional Monotonicity

Metod Jazbec
James Allingham
Dan Zhang
Eric Nalisnick

Modern predictive models are often deployed to environments in which computational budgets are dynamic. Anytime algorithms are well-suited to such environments as, at any point during computation, they can output a prediction whose quality is a function of computation time. Early-exit neural networks have garnered attention in the context of anytime computation due to their capability to provide intermediate predictions at various stages throughout the network. However, we demonstrate that current early-exit networks are not directly applicable to anytime settings, as the quality of predictions for individual data points is not guaranteed to improve with longer computation. To address this shortcoming, we propose an elegant post-hoc modification, based on the Product-of-Experts, that encourages an early-exit network to become gradually confident. This gives our deep models the property of conditional monotonicity in the prediction quality---an essential building block towards truly anytime predictive modeling using early-exit architectures. Our empirical results on standard image-classification tasks demonstrate that such behaviors can be achieved while preserving competitive accuracy on average.

PDF Details

JMLR Journal 2021 Journal Article

Normalizing Flows for Probabilistic Modeling and Inference

George Papamakarios
Eric Nalisnick
Danilo Jimenez Rezende
Shakir Mohamed
Balaji Lakshminarayanan

Normalizing flows provide a general mechanism for defining expressive probability distributions, only requiring the specification of a (usually simple) base distribution and a series of bijective transformations. There has been much recent work on normalizing flows, ranging from improving their expressive power to expanding their application. We believe the field has now matured and is in need of a unified perspective. In this review, we attempt to provide such a perspective by describing flows through the lens of probabilistic modeling and inference. We place special emphasis on the fundamental principles of flow design, and discuss foundational topics such as expressive power and computational trade-offs. We also broaden the conceptual framing of flows by relating them to more general probability transformations. Lastly, we summarize the use of flows for tasks such as generative modeling, approximate inference, and supervised learning. [abs] [ pdf ][ bib ] &copy JMLR 2021. ( edit, beta )

PDF Details

NeurIPS Conference 2019 Conference Paper

Bayesian Batch Active Learning as Sparse Subset Approximation

Robert Pinsler
Jonathan Gordon
Eric Nalisnick
José Miguel Hernández-Lobato

Leveraging the wealth of unlabeled data produced in recent years provides great potential for improving supervised models. When the cost of acquiring labels is high, probabilistic active learning methods can be used to greedily select the most informative data points to be labeled. However, for many large-scale problems standard greedy procedures become computationally infeasible and suffer from negligible model change. In this paper, we introduce a novel Bayesian batch active learning approach that mitigates these issues. Our approach is motivated by approximating the complete data posterior of the model parameters. While naive batch construction methods result in correlated queries, our algorithm produces diverse batches that enable efficient active learning at scale. We derive interpretable closed-form solutions akin to existing active learning procedures for linear models, and generalize to arbitrary models using random projections. We demonstrate the benefits of our approach on several large-scale regression and classification tasks.

PDF Details