Search Results for author:
Found 71 papers, 24 papers with code
On the Challenges and Opportunities in Generative AI
The field of deep generative modeling has grown rapidly and consistently over the years.
Robust Anomaly Detection for Particle Physics Using Multi-Background Representation Learning
Anomaly, or out-of-distribution, detection is a promising tool for aiding discoveries of new particles or processes in particle physics.
When accurate prediction models yield harmful self-fulfilling prophecies
These models are harmful self-fulfilling prophecies: their deployment harms a group of patients but the worse outcome of these patients does not invalidate the predictive power of the model.
Quantifying Impairment and Disease Severity Using AI Models Trained on Healthy Subjects
We applied the COBRA score to address a key limitation of current clinical evaluation of upper-body impairment in stroke patients.
Stochastic interpolants with data-dependent couplings
In this work, using the framework of stochastic interpolants, we formalize how to \textit{couple} the base and the target densities, whereby samples from the base are computed conditionally given samples from the target in a way that is different from (but does preclude) incorporating information about class labels or continuous embeddings.
Don't blame Dataset Shift! Shortcut Learning due to Gradients and Cross Entropy
However, even when the stable feature determines the label in the training distribution and the shortcut does not provide any additional information, like in perception tasks, default-ERM still exhibits shortcut learning.
When More is Less: Incorporating Additional Datasets Can Hurt Performance By Introducing Spurious Correlations
In machine learning, incorporating more data is often seen as a reliable strategy for improving model performance; this work challenges that notion by demonstrating that the addition of external datasets in many cases can hurt the resulting model's performance.
An Effective Meaningful Way to Evaluate Survival Models
One straightforward metric to evaluate a survival prediction model is based on the Mean Absolute Error (MAE) -- the average of the absolute difference between the time predicted by the model and the true event time, over all subjects.
A dynamic risk score for early prediction of cardiogenic shock using machine learning
We developed a deep learning-based risk stratification tool, called CShock, for patients admitted into the cardiac ICU with acute decompensated heart failure and/or myocardial infarction to predict onset of cardiogenic shock.
Don't be fooled: label leakage in explanation methods and the importance of their quantitative evaluation
Feature attribution methods identify which features of an input most influence a model's output.
Beyond Distribution Shift: Spurious Features Through the Lens of Training Dynamics
(3) We empirically show that the harmful spurious features can be detected by observing the learning dynamics of the DNN's early layers.
Where to Diffuse, How to Diffuse, and How to Get Back: Automated Learning for Multivariate Diffusions
For example, extending the inference process with auxiliary variables leads to improved sample quality.
Robustness to Spurious Correlations Improves Semantic Out-of-Distribution Detection
The detection of shared-nuisance out-of-distribution (SN-OOD) inputs is particularly relevant in real-world applications, as anomalies and in-distribution inputs tend to be captured in the same settings during deployment.
On the Feasibility of Machine Learning Augmented Magnetic Resonance for Point-of-Care Identification of Disease
We hypothesise that the disease classification task can be solved using a very small tailored subset of k-space data, compared to image reconstruction.
Nuisances via Negativa: Adjusting for Spurious Correlations via Data Augmentation
In prediction tasks, there exist features that are related to the label in the same way across different settings for that task; these are semantic features or semantics.
From algorithms to action: improving patient care requires causality
In cancer research there is much interest in building and validating outcome predicting outcomes to support treatment decisions.
Survival Mixture Density Networks
Survival MDN applies an invertible positive function to the output of Mixture Density Networks (MDNs).
DIET: Conditional independence testing with marginal dependence measures of residual information
DIET tests the marginal independence of two random variables: $F(x \mid z)$ and $F(y \mid z)$ where $F(\cdot \mid z)$ is a conditional cumulative distribution function (CDF).
Set Norm and Equivariant Skip Connections: Putting the Deep in Deep Sets
However, we show that existing permutation invariant architectures, Deep Sets and Set Transformer, can suffer from vanishing or exploding gradients when they are deep.
New-Onset Diabetes Assessment Using Artificial Intelligence-Enhanced Electrocardiography
We found ECG-based assessment outperforms the ADA Risk test, achieving a higher area under the curve (0. 80 vs. 0. 68) and positive predictive value (13% vs. 9%) -- 2. 6 times the prevalence of diabetes in the cohort.
Quantile Filtered Imitation Learning
We introduce quantile filtered imitation learning (QFIL), a novel policy improvement operator designed for offline reinforcement learning.
Learning Invariant Representations with Missing Data
Enforcing such independencies requires nuisances to be observed during training.
Inverse-Weighted Survival Games
When the loss is proper, we show that the games always have the true failure and censoring distributions as a stationary point.
FastSHAP: Real-Time Shapley Value Estimation
Shapley values are widely used to explain black-box models, but they are costly to calculate because they require many model evaluations.
Understanding Failures in Out-of-Distribution Detection with Deep Generative Models
Deep generative models (DGMs) seem a natural fit for detecting out-of-distribution (OOD) inputs, but such models have been shown to assign higher probabilities or densities to OOD images than images from the training distribution.
Out-of-Distribution Detection
Out of Distribution (OOD) Detection
Out-of-distribution Generalization in the Presence of Nuisance-Induced Spurious Correlations
NURD finds a representation from this set that is most informative of the label under the nuisance-randomized distribution, and we prove that this representation achieves the highest performance regardless of the nuisance-label relationship.
Offline RL Without Off-Policy Evaluation
In addition, we hypothesize that the strong performance of the one-step algorithm is due to a combination of favorable structure in the environment and behavior policy.
Have We Learned to Explain?: How Interpretability Methods Can Learn to Encode Predictions in their Interpretations
While the need for interpretable machine learning has been established, many common approaches are slow, lack fidelity, or hard to evaluate.
Causal Estimation with Functional Confounders
Causal inference relies on two fundamental assumptions: ignorability and positivity.
X-CAL: Explicit Calibration for Survival Analysis
A survival model's calibration can be measured using, for instance, distributional calibration (D-CALIBRATION) [Haider et al., 2020] which computes the squared difference between the observed and predicted number of events within different time intervals.
General Control Functions for Causal Effect Estimation from IVs
Causal effect estimation relies on separating the variation in the outcome into parts due to the treatment and due to the confounders.
Probabilistic Machine Learning for Healthcare
Machine learning can be used to make sense of healthcare data.
Deep Direct Likelihood Knockoffs
Predictive modeling often uses black box machine learning methods, such as deep neural networks, to achieve state-of-the-art performance.
Offline Contextual Bandits with Overparameterized Models
We show that this discrepancy is due to the \emph{action-stability} of their objectives.
The Counterfactual $χ$-GAN
Effect estimates, such as the average treatment effect (ATE), are then estimated as expectations under the reweighted or matched distribution, P .
Energy-Inspired Models: Learning with Sampler-Induced Distributions
Motivated by this, we consider the sampler-induced distribution as the model of interest and maximize the likelihood of this model.
GATO: Gates Are Not the Only Option
GATO is constructed so that part of its hidden state does not have vanishing gradients, regardless of sequence length.
Model-Agnostic Feature Selection with Additional Mutual Information
We show that f-divergences provide a broad class of proper test statistics.
Population Predictive Checks
However, PPCs use the data twice -- both to calculate the posterior predictive and to evaluate it -- which can lead to overconfident assessments of the quality of a model.
General Control Functions for Causal Effect Estimation from Instrumental Variables
Causal effect estimation relies on separating the variation in the outcome into parts due to the treatment and due to the confounders.
Reproducibility in Machine Learning for Health
Machine learning algorithms designed to characterize, monitor, and intervene on human health (ML4H) are expected to perform safely and reliably when operating at scale, potentially outside strict human supervision.
Adversarial Examples for Electrocardiograms
For this model, we utilized a new technique to generate smoothed examples to produce signals that are 1) indistinguishable to cardiologists from the original examples and 2) incorrectly classified by the neural network.
ClinicalBERT: Modeling Clinical Notes and Predicting Hospital Readmission
Clinical notes contain information about patients that goes beyond structured data like lab values and medications.
Kernelized Complete Conditional Stein Discrepancy
We introduce kernelized complete conditional Stein discrepancies (KCC-SDs).
Revisiting Auxiliary Latent Variables in Generative Models
The success of enriching the variational family with auxiliary latent variables motivates applying the same techniques to the generative model.
The Random Conditional Distribution for Higher-Order Probabilistic Inference
The need to condition distributional properties such as expectation, variance, and entropy arises in algorithmic fairness, model simplification, robustness and many other areas.
Support and Invertibility in Domain-Invariant Representations
In this work, we give generalization bounds for unsupervised domain adaptation that hold for any representation function by acknowledging the cost of non-invertibility.
The Variational Predictive Natural Gradient
Unlike traditional natural gradients for variational inference, this natural gradient accounts for the relationship between model parameters and variational parameters.
Soft Constraints for Inference with Declarative Knowledge
We develop a likelihood free inference procedure for conditioning a probabilistic model on a predicate.
A Review of Challenges and Opportunities in Machine Learning for Health
Modern electronic health records (EHRs) provide data to answer clinically meaningful questions.
Multiple Causal Inference with Latent Confounding
Together, these assumptions lead to a confounder estimator regularized by mutual information.
Noisin: Unbiased Regularization for Recurrent Neural Networks
On the Penn Treebank, the method with Noisin more quickly reaches state-of-the-art performance.
Noise-Based Regularizers for Recurrent Neural Networks
We develop a noise-based regularization method for RNNs.
Variational Inference via \chi Upper Bound Minimization
In this paper we propose CHIVI, a black-box variational inference algorithm that minimizes $D_{\chi}(p || q)$, the $\chi$-divergence from $p$ to $q$.
Variational Sequential Monte Carlo
The success of variational approaches depends on (i) formulating a flexible parametric family of distributions, and (ii) optimizing the parameters to find the member of this family that most closely approximates the exact posterior.
Proximity Variational Inference
Consequently, PVI is less sensitive to initialization and optimization quirks and finds better local optima.
Hierarchical Implicit Models and Likelihood-Free Variational Inference
Implicit probabilistic models are a flexible class of models defined by a simulation process for data.
Variational Inference via $χ$-Upper Bound Minimization
In this paper we propose CHIVI, a black-box variational inference algorithm that minimizes $D_{\chi}(p || q)$, the $\chi$-divergence from $p$ to $q$.
Operator Variational Inference
Though this divergence has been widely used, the resultant posterior approximation can suffer from undesirable statistical properties.
Deep Survival Analysis
The electronic health record (EHR) provides an unprecedented opportunity to build actionable tools to support physicians at the point of care.
The Population Posterior and Bayesian Modeling on Streams
Many modern data analysis problems involve inferences from streaming data.
The Variational Gaussian Process
Variational inference is a powerful tool for approximate inference, and it has been recently applied for representation learning with deep generative models.
Dynamic Poisson Factorization
Models for recommender systems use latent factors to explain the preferences and behaviors of users with respect to a set of items (e. g., movies, books, academic papers).
The Population Posterior and Bayesian Inference on Streams
Many modern data analysis problems involve inferences from streaming data.
Correlated Random Measures
We develop correlated random measures, random measures where the atom weights can exhibit a flexible pattern of dependence, and use them to develop powerful hierarchical Bayesian nonparametric models.
Automatic Variational Inference in Stan
With ADVI we can use variational inference on any model we write in Stan.
Deep Exponential Families
We describe \textit{deep exponential families} (DEFs), a class of latent variable models that are inspired by the hidden structures used in deep neural networks.
Variational Tempering
Lastly, we develop local variational tempering, which assigns a latent temperature to each data point; this allows for dynamic annealing that varies across data.
Black Box Variational Inference
We evaluate our method against the corresponding black box sampling based methods.
