To address the bias of the canonical two-way fixed effects estimator for difference-in-differences under staggered adoptions, Wooldridge (2021) proposed the extended two-way fixed effects estimator, which adds many parameters. However, this reduces efficiency. Restricting some of these parameters to be equal helps, but ad hoc restrictions may reintroduce bias. We propose a machine learning estimator with a single tuning parameter, fused extended two-way fixed effects (FETWFE), that enables automatic data-driven selection of these restrictions. We prove that under an appropriate sparsity assumption FETWFE identifies the correct restrictions with probability tending to one. We also prove the consistency, asymptotic normality, and oracle efficiency of FETWFE for two classes of heterogeneous marginal treatment effect estimators under either conditional or marginal parallel trends, and we prove consistency for two classes of conditional average treatment effects under conditional parallel trends. We demonstrate FETWFE in simulation studies and an empirical application.

We consider the problem of predicting how the likelihood of an outcome of interest for a patient changes over time as we observe more of the patient data. To solve this problem, we propose a supervised contrastive learning framework that learns an embedding representation for each time step of a patient time series. Our framework learns the embedding space to have the following properties: (1) nearby points in the embedding space have similar predicted class probabilities, (2) adjacent time steps of the same time series map to nearby points in the embedding space, and (3) time steps with very different raw feature vectors map to far apart regions of the embedding space. To achieve property (3), we employ a nearest neighbor pairing mechanism in the raw feature space. This mechanism also serves as an alternative to data augmentation, a key ingredient of contrastive learning, which lacks a standard procedure that is adequately realistic for clinical tabular data, to our knowledge. We demonstrate that our approach outperforms state-of-the-art baselines in predicting mortality of septic patients (MIMIC-III dataset) and tracking progression of cognitive impairment (ADNI dataset). Our method also consistently recovers the correct synthetic dataset embedding structure across experiments, a feat not achieved by baselines. Our ablation experiments show the pivotal role of our nearest neighbor pairing.

Survival Analysis (SA) constitutes the default method for time-to-event modeling due to its ability to estimate event probabilities of sparsely occurring events over time. In this work, we show how to improve the training and inference of SA models by decoupling their full expression into (1) an aggregated baseline hazard, which captures the overall behavior of a given population, and (2) independently distributed survival scores, which model idiosyncratic probabilistic dynamics of its given members, in a fully parametric setting. The proposed inference method is shown to dynamically handle right-censored observation horizons, and to achieve competitive performance when compared to other state-of-the-art methods in a variety of real-world datasets, including computationally inefficient Deep Learning-based SA methods and models that require MCMC for inference. Nevertheless, our method achieves robust results from the outset, while not being subjected to fine-tuning or hyperparameter optimization.

Inferring a diffusion equation from discretely-observed measurements is a statistical challenge of significant importance in a variety of fields, from single-molecule tracking in biophysical systems to modeling financial instruments. Assuming that the underlying dynamical process obeys a $d$-dimensional stochastic differential equation of the form $$\mathrm{d}\boldsymbol{x}_t=\boldsymbol{b}(\boldsymbol{x}_t)\mathrm{d} t+\Sigma(\boldsymbol{x}_t)\mathrm{d}\boldsymbol{w}_t,$$ we propose neural network-based estimators of both the drift $\boldsymbol{b}$ and the spatially-inhomogeneous diffusion tensor $D = \Sigma\Sigma^{T}$ and provide statistical convergence guarantees when $\boldsymbol{b}$ and $D$ are $s$-H\"older continuous. Notably, our bound aligns with the minimax optimal rate $N^{-\frac{2s}{2s+d}}$ for nonparametric function estimation even in the presence of correlation within observational data, which necessitates careful handling when establishing fast-rate generalization bounds. Our theoretical results are bolstered by numerical experiments demonstrating accurate inference of spatially-inhomogeneous diffusion tensors.

Quantifying variable importance is essential for answering high-stakes questions in fields like genetics, public policy, and medicine. Current methods generally calculate variable importance for a given model trained on a given dataset. However, for a given dataset, there may be many models that explain the target outcome equally well; without accounting for all possible explanations, different researchers may arrive at many conflicting yet equally valid conclusions given the same data. Additionally, even when accounting for all possible explanations for a given dataset, these insights may not generalize because not all good explanations are stable across reasonable data perturbations. We propose a new variable importance framework that quantifies the importance of a variable across the set of all good models and is stable across the data distribution. Our framework is extremely flexible and can be integrated with most existing model classes and global variable importance metrics. We demonstrate through experiments that our framework recovers variable importance rankings for complex simulation setups where other methods fail. Further, we show that our framework accurately estimates the true importance of a variable for the underlying data distribution. We provide theoretical guarantees on the consistency and finite sample error rates for our estimator. Finally, we demonstrate its utility with a real-world case study exploring which genes are important for predicting HIV load in persons with HIV, highlighting an important gene that has not previously been studied in connection with HIV. Code is available at https://github.com/jdonnelly36/Rashomon_Importance_Distribution.

Sharpness-aware minimization (SAM) and related adversarial deep-learning methods can drastically improve generalization, but their underlying mechanisms are not yet fully understood. Here, we establish SAM as a relaxation of the Bayes objective where the expected negative-loss is replaced by the optimal convex lower bound, obtained by using the so-called Fenchel biconjugate. The connection enables a new Adam-like extension of SAM to automatically obtain reasonable uncertainty estimates, while sometimes also improving its accuracy. By connecting adversarial and Bayesian methods, our work opens a new path to robustness.

We prove a few representer theorems for a localised version of the regularised and multiview support vector machine learning problem introduced by H.Q.~Minh, L.~Bazzani, and V.~Murino, \textit{Journal of Machine Learning Research}, \textbf{17}(2016) 1--72, that involves operator valued positive semidefinite kernels and their reproducing kernel Hilbert spaces. The results concern general cases when convex or nonconvex loss functions and finite or infinite dimensional input spaces are considered. We show that the general framework allows infinite dimensional input spaces and nonconvex loss functions for some special cases, in particular in case the loss functions are G\^ateaux differentiable. Detailed calculations are provided for the exponential least squares loss functions that leads to partially nonlinear problems.

Path signatures have been proposed as a powerful representation of paths that efficiently captures the path's analytic and geometric characteristics, having useful algebraic properties including fast concatenation of paths through tensor products. Signatures have recently been widely adopted in machine learning problems for time series analysis. In this work we establish connections between value functions typically used in optimal control and intriguing properties of path signatures. These connections motivate our novel control framework with signature transforms that efficiently generalizes the Bellman equation to the space of trajectories. We analyze the properties and advantages of the framework, termed signature control. In particular, we demonstrate that (i) it can naturally deal with varying/adaptive time steps; (ii) it propagates higher-level information more efficiently than value function updates; (iii) it is robust to dynamical system misspecification over long rollouts. As a specific case of our framework, we devise a model predictive control method for path tracking. This method generalizes integral control, being suitable for problems with unknown disturbances. The proposed algorithms are tested in simulation, with differentiable physics models including typical control and robotics tasks such as point-mass, curve following for an ant model, and a robotic manipulator.

We propose a conditional stochastic interpolation (CSI) approach to learning conditional distributions. CSI learns probability flow equations or stochastic differential equations that transport a reference distribution to the target conditional distribution. This is achieved by first learning the drift function and the conditional score function based on conditional stochastic interpolation, which are then used to construct a deterministic process governed by an ordinary differential equation or a diffusion process for conditional sampling. In our proposed CSI model, we incorporate an adaptive diffusion term to address the instability issues arising during the training process. We provide explicit forms of the conditional score function and the drift function in terms of conditional expectations under mild conditions, which naturally lead to an nonparametric regression approach to estimating these functions. Furthermore, we establish non-asymptotic error bounds for learning the target conditional distribution via conditional stochastic interpolation in terms of KL divergence, taking into account the neural network approximation error. We illustrate the application of CSI on image generation using a benchmark image dataset.

Both models demonstrate mastery of the pre-training task but neither consistently outperform XGBoost on downstream supervised tasks. We encourage future work to focus on joint modeling of multiple EHR data categories and to include tree-based baselines in their evaluations. In recent years, self-supervised pre-training of masked language models (MLMs) (see Appendix A for background) has demonstrated remarkable success across a wide range of machine learning problems and has led to significant downstream improvements across diverse tasks in natural language processing (Liu et al., 2019; Devlin et al., 2019; Raffel et al., 2020). There is considerable excitement surrounding the potential of large pre-trained MLMs to achieve similar success in medical applications. For instance, existing applications of MLMs in medicine have already yielded promising results in tasks related to medical text understanding (Lee et al., 2020; Alsentzer et al., 2019; Huang et al., 2019; Yang et al., 2019; Beltagy et al., 2019). Laboratory data is abundant, routinely collected, less biased compared to other types of data in electronic health records (EHRs) like billing codes (Beam et al., 2021), and directly measure a patient's physiological state, offering a valuable opportunity for creating a medical foundation model. However, there is a large body of evidence showing that deep learning is consistently outperformed on so-called "tabular" data prediction tasks by traditional machine learning techniques like random forests, XGBoost, and even simple regression models (Bellamy et al., 2020; Finlayson et al., 2023; Sharma, 2013). The reasons for this are only partially understood, but previous work (Grinsztajn et al., 2022) has suggested that this phenomenon may be caused by a rotational invariance in deep learning models that is harmful for tabular data. More broadly, the success of deep learning is thought to be largely due to inductive biases that can be leveraged for images, text, and graphs. These inductive biases are absent or only weakly present in tabular data. Conversely, tree-based methods are scale invariant and robust to uninformative features. We evaluated both models on several downstream outcome prediction tasks and validated the success of pre-training with a set of intrinsic evaluations.