Imitation learning has proven to be a powerful tool for training complex visuomotor policies. However, current methods often require hundreds to thousands of expert demonstrations to handle high-dimensional visual observations. A key reason for this poor data efficiency is that visual representations are predominantly either pretrained on out-of-domain data or trained directly through a behavior cloning objective.

A.1 Data, Models, and Model Accuracies Images are pixel-wise normalized by the mean and standard deviation of the training images for each dataset, and for ImageNet all images are center cropped and resized to 224 224; this preprocessing is done before any interpolating paths are constructed. Our implementations are based on Cubuk et al. [6]; we use the same optimizer (stochastic gradient descent with momentum) and cosine learning rate schedule. We train without data augmentation (to ensure all models are trained on exactly the same examples), except for experiments that explicitly vary data augmentation. Without data augmentation, the test accuracies of our models are shown in Table 1. The top-1 classification accuracies of these models are presented in Table 2. Model ImageNet Test Accuracy (%) A.2 Linear Interpolation: Methodological Details For each sampled path, we compute the discrete Fourier transform (DFT) of the prediction function along the path separately for each of the M class predictions, take the (real) magnitude of the resulting (complex) DFT coefficients, and average them among the M classes.

Sign stochastic gradient descent (signSGD) is a communication-efficient method that transmits only the sign of stochastic gradients for parameter updating.

In this paper, we study stochastic structured bandits for minimizing regret. The fact that the popular optimistic algorithms do not achieve the asymptotic instancedependent regret optimality (asymptotic optimality for short) has recently allured researchers. On the other hand, it is known that one can achieve a bounded regret (i.e., does not grow indefinitely with n) in certain instances. Unfortunately, existing asymptotically optimal algorithms rely on forced sampling that introduces an ω(1) term w.r.t. the time horizon n in their regret, failing to adapt to the "easiness" of the instance. In this paper, we focus on the finite hypothesis class and ask if one can achieve the asymptotic optimality while enjoying bounded regret whenever possible. We provide a positive answer by introducing a new algorithm called CRush Optimism with Pessimism (CROP) that eliminates optimistic hypotheses by pulling the informative arms indicated by a pessimistic hypothesis.

Modern machine learning models deployed often encounter distribution shifts in real-world applications, manifesting as covariate or semantic out-of-distribution (OOD) shifts. These shifts give rise to challenges in OOD generalization and OOD detection. This paper introduces a novel, integrated approach AHA (Adaptive Human-Assisted OOD learning) to simultaneously address both OOD generalization and detection through a human-assisted framework by labeling data in the wild. Our approach strategically labels examples within a novel maximum disambiguation region, where the number of semantic and covariate OOD data roughly equalizes. By labeling within this region, we can maximally disambiguate the two types of OOD data, thereby maximizing the utility of the fixed labeling budget. Our algorithm first utilizes a noisy binary search algorithm that identifies the maximal disambiguation region with high probability. The algorithm then continues with annotating inside the identified labeling region, reaping the full benefit of human feedback.

Several challenges make it difficult for sparse neural networks to compete with dense models.

We seek an entropy estimator for discrete distributions with fully empirical accuracy bounds. As stated, this goal is infeasible without some prior assumptions on the distribution. We discover that a certain information moment assumption renders the problem feasible. We argue that the moment assumption is natural and, in some sense, minimalistic -- weaker than finite support or tail decay conditions. Under the moment assumption, we provide the first finite-sample entropy estimates for infinite alphabets, nearly recovering the known minimax rates. Moreover, we demonstrate that our empirical bounds are significantly sharper than the state-ofthe-art bounds, for various natural distributions and non-trivial sample regimes. Along the way, we give a dimension-free analogue of the Cover-Thomas result on entropy continuity (with respect to total variation distance) for finite alphabets, which may be of independent interest.

Understanding how the collective activity of neural populations relates to computation and ultimately behavior is a key goal in neuroscience. To this end, statistical methods which describe high-dimensional neural time series in terms of lowdimensional latent dynamics have played a fundamental role in characterizing neural systems. Yet, what constitutes a successful method involves two opposing criteria: (1) methods should be expressive enough to capture complex nonlinear dynamics, and (2) they should maintain a notion of interpretability often only warranted by simpler linear models. In this paper, we develop an approach that balances these two objectives: the Gaussian Process Switching Linear Dynamical System (gpSLDS). Our method builds on previous work modeling the latent state evolution via a stochastic differential equation whose nonlinear dynamics are described by a Gaussian process (GP-SDEs). We propose a novel kernel function which enforces smoothly interpolated locally linear dynamics, and therefore expresses flexible - yet interpretable - dynamics akin to those of recurrent switching linear dynamical systems (rSLDS).