Zhen Lin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhen Lin has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Jimeng Sun
Department of Computer Science, University of Illinois Urbana-Champaign, Champaign, USA
8 shared publications - Shubhendu Trivedi
New Jersey Institute of Technology
5 shared publications - Brian Nord
Kavli Institute for Cosmological Physics, University of Chicago, Chicago, IL 60637
2 shared publications - Camille Avestruz
University of Michigan
2 shared publications Cao Xiao
2 shared publications- Benjamin P. Danek
Center for Alzheimer’s and Related Dementias, National Institute on Aging, National Institutes of Health, Bethesda, MD, 20892, USA
1 shared publication
Latest Publications
- Top Cited PapersOpen Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 6107-6117
Uncertainty Quantification and Confidence Calibration in Large Language Models: A Survey
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 6107-6117
Top Cited PapersOpen AccessUncertainty Quantification and Confidence Calibration in Large Language Models: A Survey
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 10351-10368
Contextualized Sequence Likelihood: Enhanced Confidence Scores for Natural Language Generation
Zhen Lin, Shubhendu Trivedi, Jimeng Sun. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 10351-10368
Open AccessContextualized Sequence Likelihood: Enhanced Confidence Scores for Natural Language Generation
Zhen Lin, Shubhendu Trivedi, Jimeng Sun. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- Open Access
Conformal Drug Property Prediction with Density Estimation under Covariate Shift
In drug discovery, it is vital to confirm the predictions of pharmaceutical properties from computational models using costly wet-lab experiments. Hence, obtaining reliable uncertainty estimates is crucial for prioritizing drug molecules for subsequent experimental validation. Conformal Prediction (CP) is a promising tool for creating such prediction sets for molecular properties with a coverage guarantee. However, the exchangeability assumption of CP is often challenged with covariate shift in drug discovery tasks: Most datasets contain limited labeled data, which may not be representative of the vast chemical space from which molecules are drawn. To address this limitation, we propose a method called CoDrug that employs an energy-based model leveraging both training data and unlabelled data, and Kernel Density Estimation (KDE) to assess the densities of a molecule set. The estimated densities are then used to weigh the molecule samples while building prediction sets and rectifying for distribution shift. In extensive experiments involving realistic distribution drifts in various small-molecule drug discovery tasks, we demonstrate the ability of CoDrug to provide valid prediction sets and its utility in addressing the distribution shift arising from de novo drug design models. On average, using CoDrug can reduce the coverage gap by over 35% when compared to conformal prediction sets not adjusted for covariate shift.
Open AccessConformal Drug Property Prediction with Density Estimation under Covariate Shift
In drug discovery, it is vital to confirm the predictions of pharmaceutical properties from computational models using costly wet-lab experiments. Hence, obtaining reliable uncertainty estimates is crucial for prioritizing drug molecules for subsequent experimental validation. Conformal Prediction (CP) is a promising tool for creating such prediction sets for molecular properties with a coverage guarantee. However, the exchangeability assumption of CP is often challenged with covariate shift in drug discovery tasks: Most datasets contain limited labeled data, which may not be representative of the vast chemical space from which molecules are drawn. To address this limitation, we propose a method called CoDrug that employs an energy-based model leveraging both training data and unlabelled data, and Kernel Density Estimation (KDE) to assess the densities of a molecule set. The estimated densities are then used to weigh the molecule samples while building prediction sets and rectifying for distribution shift. In extensive experiments involving realistic distribution drifts in various small-molecule drug discovery tasks, we demonstrate the ability of CoDrug to provide valid prediction sets and its utility in addressing the distribution shift arising from de novo drug design models. On average, using CoDrug can reduce the coverage gap by over 35% when compared to conformal prediction sets not adjusted for covariate shift.
- Open Access
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5788-5789
PyHealth: A Deep Learning Toolkit for Healthcare Applications
Deep learning (DL) has emerged as a promising tool in healthcare applications. However, the reproducibility of many studies in this field is limited by the lack of accessible code implementations and standard benchmarks. To address the issue, we create PyHealth, a comprehensive library to build, deploy, and validate DL pipelines for healthcare applications. PyHealth supports various data modalities, including electronic health records (EHRs), physiological signals, medical images, and clinical text. It offers various advanced DL models and maintains comprehensive medical knowledge systems. The library is designed to support both DL researchers and clinical data scientists. Upon the time of writing, PyHealth has received 633 stars, 130 forks, and 15k+ downloads in total on GitHub. This tutorial will provide an overview of PyHealth, present different modules, and showcase their functionality through hands-on demos. Participants can follow along and gain hands-on experience on the Google Colab platform during the session.
- 4 August 2023
- Published by Association for Computing Machinery (ACM)
- p. 5788-5789
Open AccessPyHealth: A Deep Learning Toolkit for Healthcare Applications
Deep learning (DL) has emerged as a promising tool in healthcare applications. However, the reproducibility of many studies in this field is limited by the lack of accessible code implementations and standard benchmarks. To address the issue, we create PyHealth, a comprehensive library to build, deploy, and validate DL pipelines for healthcare applications. PyHealth supports various data modalities, including electronic health records (EHRs), physiological signals, medical images, and clinical text. It offers various advanced DL models and maintains comprehensive medical knowledge systems. The library is designed to support both DL researchers and clinical data scientists. Upon the time of writing, PyHealth has received 633 stars, 130 forks, and 15k+ downloads in total on GitHub. This tutorial will provide an overview of PyHealth, present different modules, and showcase their functionality through hands-on demos. Participants can follow along and gain hands-on experience on the Google Colab platform during the session.
- Open Access
Fast Online Value-Maximizing Prediction Sets with Conformal Cost Control
Many real-world multi-label prediction problems involve set-valued predictions that must satisfy specific requirements dictated by downstream usage. We focus on a typical scenario where such requirements, separately encoding $\textit{value}$ and $\textit{cost}$, compete with each other. For instance, a hospital might expect a smart diagnosis system to capture as many severe, often co-morbid, diseases as possible (the value), while maintaining strict control over incorrect predictions (the cost). We present a general pipeline, dubbed as FavMac, to maximize the value while controlling the cost in such scenarios. FavMac can be combined with almost any multi-label classifier, affording distribution-free theoretical guarantees on cost control. Moreover, unlike prior works, it can handle real-world large-scale applications via a carefully designed online update mechanism, which is of independent interest. Our methodological and theoretical contributions are supported by experiments on several healthcare tasks and synthetic datasets - FavMac furnishes higher value compared with several variants and baselines while maintaining strict cost control. Our code is available at https://github.com/zlin7/FavMacOpen AccessFast Online Value-Maximizing Prediction Sets with Conformal Cost Control
Many real-world multi-label prediction problems involve set-valued predictions that must satisfy specific requirements dictated by downstream usage. We focus on a typical scenario where such requirements, separately encoding $\textit{value}$ and $\textit{cost}$, compete with each other. For instance, a hospital might expect a smart diagnosis system to capture as many severe, often co-morbid, diseases as possible (the value), while maintaining strict control over incorrect predictions (the cost). We present a general pipeline, dubbed as FavMac, to maximize the value while controlling the cost in such scenarios. FavMac can be combined with almost any multi-label classifier, affording distribution-free theoretical guarantees on cost control. Moreover, unlike prior works, it can handle real-world large-scale applications via a carefully designed online update mechanism, which is of independent interest. Our methodological and theoretical contributions are supported by experiments on several healthcare tasks and synthetic datasets - FavMac furnishes higher value compared with several variants and baselines while maintaining strict cost control. Our code is available at https://github.com/zlin7/FavMac - Open Access
Taking a Step Back with KCal: Multi-Class Kernel-Based Calibration for Deep Neural Networks
Deep neural network (DNN) classifiers are often overconfident, producing miscalibrated class probabilities. In high-risk applications like healthcare, practitioners require $\textit{fully calibrated}$ probability predictions for decision-making. That is, conditioned on the prediction $\textit{vector}$, $\textit{every}$ class' probability should be close to the predicted value. Most existing calibration methods either lack theoretical guarantees for producing calibrated outputs, reduce classification accuracy in the process, or only calibrate the predicted class. This paper proposes a new Kernel-based calibration method called KCal. Unlike existing calibration procedures, KCal does not operate directly on the logits or softmax outputs of the DNN. Instead, KCal learns a metric space on the penultimate-layer latent embedding and generates predictions using kernel density estimates on a calibration set. We first analyze KCal theoretically, showing that it enjoys a provable $\textit{full}$ calibration guarantee. Then, through extensive experiments across a variety of datasets, we show that KCal consistently outperforms baselines as measured by the calibration error and by proper scoring rules like the Brier Score.Open AccessTaking a Step Back with KCal: Multi-Class Kernel-Based Calibration for Deep Neural Networks
Deep neural network (DNN) classifiers are often overconfident, producing miscalibrated class probabilities. In high-risk applications like healthcare, practitioners require $\textit{fully calibrated}$ probability predictions for decision-making. That is, conditioned on the prediction $\textit{vector}$, $\textit{every}$ class' probability should be close to the predicted value. Most existing calibration methods either lack theoretical guarantees for producing calibrated outputs, reduce classification accuracy in the process, or only calibrate the predicted class. This paper proposes a new Kernel-based calibration method called KCal. Unlike existing calibration procedures, KCal does not operate directly on the logits or softmax outputs of the DNN. Instead, KCal learns a metric space on the penultimate-layer latent embedding and generates predictions using kernel density estimates on a calibration set. We first analyze KCal theoretically, showing that it enjoys a provable $\textit{full}$ calibration guarantee. Then, through extensive experiments across a variety of datasets, we show that KCal consistently outperforms baselines as measured by the calibration error and by proper scoring rules like the Brier Score. - Open Access
Conformal Prediction Intervals with Temporal Dependence
Cross-sectional prediction is common in many domains such as healthcare, including forecasting tasks using electronic health records, where different patients form a cross-section. We focus on the task of constructing valid prediction intervals (PIs) in time series regression with a cross-section. A prediction interval is considered valid if it covers the true response with (a pre-specified) high probability. We first distinguish between two notions of validity in such a setting: cross-sectional and longitudinal. Cross-sectional validity is concerned with validity across the cross-section of the time series data, while longitudinal validity accounts for the temporal dimension. Coverage guarantees along both these dimensions are ideally desirable; however, we show that distribution-free longitudinal validity is theoretically impossible. Despite this limitation, we propose Conformal Prediction with Temporal Dependence (CPTD), a procedure that is able to maintain strict cross-sectional validity while improving longitudinal coverage. CPTD is post-hoc and light-weight, and can easily be used in conjunction with any prediction model as long as a calibration set is available. We focus on neural networks due to their ability to model complicated data such as diagnosis codes for time series regression, and perform extensive experimental validation to verify the efficacy of our approach. We find that CPTD outperforms baselines on a variety of datasets by improving longitudinal coverage and often providing more efficient (narrower) PIs.
Open AccessConformal Prediction Intervals with Temporal Dependence
Cross-sectional prediction is common in many domains such as healthcare, including forecasting tasks using electronic health records, where different patients form a cross-section. We focus on the task of constructing valid prediction intervals (PIs) in time series regression with a cross-section. A prediction interval is considered valid if it covers the true response with (a pre-specified) high probability. We first distinguish between two notions of validity in such a setting: cross-sectional and longitudinal. Cross-sectional validity is concerned with validity across the cross-section of the time series data, while longitudinal validity accounts for the temporal dimension. Coverage guarantees along both these dimensions are ideally desirable; however, we show that distribution-free longitudinal validity is theoretically impossible. Despite this limitation, we propose Conformal Prediction with Temporal Dependence (CPTD), a procedure that is able to maintain strict cross-sectional validity while improving longitudinal coverage. CPTD is post-hoc and light-weight, and can easily be used in conjunction with any prediction model as long as a calibration set is available. We focus on neural networks due to their ability to model complicated data such as diagnosis codes for time series regression, and perform extensive experimental validation to verify the efficacy of our approach. We find that CPTD outperforms baselines on a variety of datasets by improving longitudinal coverage and often providing more efficient (narrower) PIs.
- 28 June 2022
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 36 (7) , 7497-7505
SCRIB: Set-Classifier with Class-Specific Risk Bounds for Blackbox Models
Despite deep learning (DL) success in classification problems, DL classifiers do not provide a sound mechanism to decide when to refrain from predicting. Recent works tried to control the overall prediction risk with classification with rejection options. However, existing works overlook the different significance of different classes. We introduce Set-classifier with class-specific RIsk Bounds (SCRIB) to tackle this problem, assigning multiple labels to each example. Given the output of a black-box model on the validation set, SCRIB constructs a set-classifier that controls the class-specific prediction risks. The key idea is to reject when the set classifier returns more than one label. We validated SCRIB on several medical applications, including sleep staging on electroencephalogram(EEG) data, X-ray COVID image classification, and atrial fibrillation detection based on electrocardiogram (ECG) data.SCRIB obtained desirable class-specific risks, which are 35%-88% closer to the target risks than baseline methods.
- 28 June 2022
- Published by Association for the Advancement of Artificial Intelligence (AAAI) in Proceedings of the AAAI Conference on Artificial Intelligence
- Vol. 36 (7) , 7497-7505
SCRIB: Set-Classifier with Class-Specific Risk Bounds for Blackbox Models
Despite deep learning (DL) success in classification problems, DL classifiers do not provide a sound mechanism to decide when to refrain from predicting. Recent works tried to control the overall prediction risk with classification with rejection options. However, existing works overlook the different significance of different classes. We introduce Set-classifier with class-specific RIsk Bounds (SCRIB) to tackle this problem, assigning multiple labels to each example. Given the output of a black-box model on the validation set, SCRIB constructs a set-classifier that controls the class-specific prediction risks. The key idea is to reject when the set classifier returns more than one label. We validated SCRIB on several medical applications, including sleep staging on electroencephalogram(EEG) data, X-ray COVID image classification, and atrial fibrillation detection based on electrocardiogram (ECG) data.SCRIB obtained desirable class-specific risks, which are 35%-88% closer to the target risks than baseline methods.
- Open Access
Locally Valid and Discriminative Prediction Intervals for Deep Learning Models
Crucial for building trust in deep learning models for critical real-world applications is efficient and theoretically sound uncertainty quantification, a task that continues to be challenging. Useful uncertainty information is expected to have two key properties: It should be valid (guaranteeing coverage) and discriminative (more uncertain when the expected risk is high). Moreover, when combined with deep learning (DL) methods, it should be scalable and affect the DL model performance minimally. Most existing Bayesian methods lack frequentist coverage guarantees and usually affect model performance. The few available frequentist methods are rarely discriminative and/or violate coverage guarantees due to unrealistic assumptions. Moreover, many methods are expensive or require substantial modifications to the base neural network. Building upon recent advances in conformal prediction [13, 33] and leveraging the classical idea of kernel regression, we propose Locally Valid and Discriminative prediction intervals (LVD), a simple, efficient, and lightweight method to construct discriminative prediction intervals (PIs) for almost any DL model. With no assumptions on the data distribution, such PIs also offer finite-sample local coverage guarantees (contrasted to the simpler marginal coverage). We empirically verify, using diverse datasets, that besides being the only locally valid method for DL, LVD also exceeds or matches the performance (including coverage rate and prediction accuracy) of existing uncertainty quantification methods, while offering additional benefits in scalability and flexibility.
Open AccessLocally Valid and Discriminative Prediction Intervals for Deep Learning Models
Crucial for building trust in deep learning models for critical real-world applications is efficient and theoretically sound uncertainty quantification, a task that continues to be challenging. Useful uncertainty information is expected to have two key properties: It should be valid (guaranteeing coverage) and discriminative (more uncertain when the expected risk is high). Moreover, when combined with deep learning (DL) methods, it should be scalable and affect the DL model performance minimally. Most existing Bayesian methods lack frequentist coverage guarantees and usually affect model performance. The few available frequentist methods are rarely discriminative and/or violate coverage guarantees due to unrealistic assumptions. Moreover, many methods are expensive or require substantial modifications to the base neural network. Building upon recent advances in conformal prediction [13, 33] and leveraging the classical idea of kernel regression, we propose Locally Valid and Discriminative prediction intervals (LVD), a simple, efficient, and lightweight method to construct discriminative prediction intervals (PIs) for almost any DL model. With no assumptions on the data distribution, such PIs also offer finite-sample local coverage guarantees (contrasted to the simpler marginal coverage). We empirically verify, using diverse datasets, that besides being the only locally valid method for DL, LVD also exceeds or matches the performance (including coverage rate and prediction accuracy) of existing uncertainty quantification methods, while offering additional benefits in scalability and flexibility.
- Open Access
- 5 August 2021
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 507 (3) , 4149-4164
DeepSZ: identification of Sunyaev–Zel’dovich galaxy clusters using deep learning
Galaxy clusters identified via the Sunyaev–Zel’dovich (SZ) effect are a key ingredient in multiwavelength cluster cosmology. We present and compare three methods of cluster identification: the standard matched filter (MF) method in SZ cluster finding, a convolutional neural networks (CNN), and a ‘combined’ identifier. We apply the methods to simulated millimeter maps for several observing frequencies for a survey similar to SPT-3G, the third-generation camera for the South Pole Telescope. The MF requires image pre-processing to remove point sources and a model for the noise, while the CNN requires very little pre-processing of images. Additionally, the CNN requires tuning of hyperparameters in the model and takes cut-out images of the sky as input, identifying the cut-out as cluster-containing or not. We compare differences in purity and completeness. The MF signal-to-noise ratio depends on both mass and redshift. Our CNN, trained for a given mass threshold, captures a different set of clusters than the MF, some with signal-to-noise-ratio below the MF detection threshold. However, the CNN tends to mis-classify cut-out whose clusters are located near the edge of the cut-out, which can be mitigated with staggered cut-out. We leverage the complementarity of the two methods, combining the scores from each method for identification. The purity and completeness are both 0.61 for MF, and 0.59 and 0.61 for CNN. The combined method yields 0.60 and 0.77, a significant increase for completeness with a modest decrease in purity. We advocate for combined methods that increase the confidence of many low signal-to-noise clusters.
- 5 August 2021
- Published by Oxford University Press (OUP) in Monthly Notices of the Royal Astronomical Society
- Vol. 507 (3) , 4149-4164
Open AccessDeepSZ: identification of Sunyaev–Zel’dovich galaxy clusters using deep learning
Galaxy clusters identified via the Sunyaev–Zel’dovich (SZ) effect are a key ingredient in multiwavelength cluster cosmology. We present and compare three methods of cluster identification: the standard matched filter (MF) method in SZ cluster finding, a convolutional neural networks (CNN), and a ‘combined’ identifier. We apply the methods to simulated millimeter maps for several observing frequencies for a survey similar to SPT-3G, the third-generation camera for the South Pole Telescope. The MF requires image pre-processing to remove point sources and a model for the noise, while the CNN requires very little pre-processing of images. Additionally, the CNN requires tuning of hyperparameters in the model and takes cut-out images of the sky as input, identifying the cut-out as cluster-containing or not. We compare differences in purity and completeness. The MF signal-to-noise ratio depends on both mass and redshift. Our CNN, trained for a given mass threshold, captures a different set of clusters than the MF, some with signal-to-noise-ratio below the MF detection threshold. However, the CNN tends to mis-classify cut-out whose clusters are located near the edge of the cut-out, which can be mitigated with staggered cut-out. We leverage the complementarity of the two methods, combining the scores from each method for identification. The purity and completeness are both 0.61 for MF, and 0.59 and 0.61 for CNN. The combined method yields 0.60 and 0.77, a significant increase for completeness with a modest decrease in purity. We advocate for combined methods that increase the confidence of many low signal-to-noise clusters.