Multi-label learning predicts a subset of labels from a given label set for an unseen instance while considering label correlations. A known challenge with multi-label classification is the long-tailed distribution of labels. Many studies focus on improving the overall predictions of the model and thus do not prioritise tail-end labels. Improving the tail-end label predictions in multi-label classifications of medical text enables the potential to understand patients better and improve care. The knowledge gained by one or more infrequent labels can impact the cause of medical decisions and treatment plans. This research presents variations of concatenated domain-specific language models, including multi-BioMed-Transformers, to achieve two primary goals. First, to improve F1 scores of infrequent labels across multi-label problems, especially with long-tail labels; second, to handle long medical text and multi-sourced electronic health records (EHRs), a challenging task for standard transformers designed to work on short input sequences. A vital contribution of this research is new state-of-the-art (SOTA) results obtained using TransformerXL for predicting medical codes. A variety of experiments are performed on the Medical Information Mart for Intensive Care (MIMIC-III) database. Results show that concatenated BioMed-Transformers outperform standard transformers in terms of overall micro and macro F1 scores and individual F1 scores of tail-end labels, while incurring lower training times than existing transformer-based solutions for long input sequences.

For weeks, Jay Deitcher has been on the hunt for a specific Miles Morales: Spider-Man toy from Spidey and His Amazing Friends. "The thing that makes the toy special is Miles's mask flips up to show his face," Deitcher says. "My son is Black, and it would be great to have a Spider-Man figure that looks like him." But even though the father of two from Albany, New York, started shopping for Hanukkah earlier than usual, he has yet to track down the elusive toy, which is sold out at many retailers. "We were already expecting a shortage, so we got him most of his other presents," he says.

Almost 300 Royal Australian Navy (RAN) members have completed officer training in 2021 – the largest cohort of new officers to graduate in a single year since the 1950's. This week, 125 RAN members graduated from the New Entry Officers' Course (NEOC) at the Royal Australian Naval College in Jervis Bay. Another 173 officers completed the world-class leadership course in the first half of 2021. This year's NEOC graduates come from all over Australia and from a range of backgrounds, including high-school leavers and professionals looking to switch careers. Chief of Navy, Vice Admiral Michael Noonan said it was great to see such a large number of new officers from diverse backgrounds graduating at a time of significant growth and change for Navy.

A comprehensive understanding of the behaviours of the various geophysical processes requires, among others, detailed investigations across temporal scales. In this work, we propose a new time series feature compilation for advancing and enriching such investigations in a hydroclimatic context. This specific compilation can facilitate largely interpretable feature investigations and comparisons in terms of temporal dependence, temporal variation, "forecastability", lumpiness, stability, nonlinearity (and linearity), trends, spikiness, curvature and seasonality. Detailed quantifications and multifaceted characterizations are herein obtained by computing the values of the proposed feature compilation across nine temporal resolutions (i.e., the 1-day, 2-day, 3-day, 7-day, 0.5-month, 1-month, 2-month, 3-month and 6-month ones) and three hydroclimatic time series types (i.e., temperature, precipitation and streamflow) for 34-year-long time series records originating from 511 geographical locations across the continental United States. Based on the acquired information and knowledge, similarities and differences between the examined time series types with respect to the evolution patterns characterizing their feature values with increasing (or decreasing) temporal resolution are identified. To our view, the similarities in these patterns are rather surprising. We also find that the spatial patterns emerging from feature-based time series clustering are largely analogous across temporal scales, and compare the features with respect to their usefulness in clustering the time series at the various temporal resolutions. For most of the features, this usefulness can vary to a notable degree across temporal resolutions and time series types, thereby pointing out the need for conducting multifaceted time series characterizations for the study of hydroclimatic similarity.

We study the problem of large-scale network embedding, which aims to learn low-dimensional latent representations for network mining applications. Recent research in the field of network embedding has led to significant progress such as DeepWalk, LINE, NetMF, NetSMF. However, the huge size of many real-world networks makes it computationally expensive to learn network embedding from the entire network. In this work, we present a novel network embedding method called "NES", which learns network embedding from a small representative subgraph. NES leverages theories from graph sampling to efficiently construct representative subgraph with smaller size which can be used to make inferences about the full network, enabling significantly improved efficiency in embedding learning. Then, NES computes the network embedding from this representative subgraph, efficiently. Compared with well-known methods, extensive experiments on networks of various scales and types demonstrate that NES achieves comparable performance and significant efficiency superiority.

Within the last few years, there has been a move towards using statistical models in conjunction with neural networks with the end goal of being able to better answer the question, "what do our models know?". From this trend, classical metrics such as Prediction Interval Coverage Probability (PICP) and new metrics such as calibration error have entered the general repertoire of model evaluation in order to gain better insight into how the uncertainty of our model compares to reality. One important component of uncertainty modeling is model uncertainty (epistemic uncertainty), a measurement of what the model does and does not know. However, current evaluation techniques tends to conflate model uncertainty with aleatoric uncertainty (irreducible error), leading to incorrect conclusions. In this paper, using posterior predictive checks, we show how calibration error and its variants are almost always incorrect to use given model uncertainty, and further show how this mistake can lead to trust in bad models and mistrust in good models. Though posterior predictive checks has often been used for in-sample evaluation of Bayesian models, we show it still has an important place in the modern deep learning world.

There is often a trade-off between building deep learning systems that are expressive enough to capture the nuances of the reality, and having the right inductive biases for efficient learning. We introduce Residual Pathway Priors (RPPs) as a method for converting hard architectural constraints into soft priors, guiding models towards structured solutions, while retaining the ability to capture additional complexity. Using RPPs, we construct neural network priors with inductive biases for equivariances, but without limiting flexibility. We show that RPPs are resilient to approximate or misspecified symmetries, and are as effective as fully constrained models even when symmetries are exact. We showcase the broad applicability of RPPs with dynamical systems, tabular data, and reinforcement learning. In Mujoco locomotion tasks, where contact forces and directional rewards violate strict equivariance assumptions, the RPP outperforms baseline model-free RL agents, and also improves the learned transition models for model-based RL.

We investigate the generalization properties of a self-training algorithm with halfspaces. The approach learns a list of halfspaces iteratively from labeled and unlabeled training data, in which each iteration consists of two steps: exploration and pruning. In the exploration phase, the halfspace is found sequentially by maximizing the unsigned-margin among unlabeled examples and then assigning pseudo-labels to those that have a distance higher than the current threshold. The pseudo-labeled examples are then added to the training set, and a new classifier is learned. This process is repeated until no more unlabeled examples remain for pseudo-labeling. In the pruning phase, pseudo-labeled samples that have a distance to the last halfspace greater than the associated unsigned-margin are then discarded. We prove that the misclassification error of the resulting sequence of classifiers is bounded and show that the resulting semi-supervised approach never degrades performance compared to the classifier learned using only the initial labeled training set. Experiments carried out on a variety of benchmarks demonstrate the efficiency of the proposed approach compared to state-of-the-art methods.

We consider a multi-armed bandit problem specified by a set of one-dimensional family exponential distributions endowed with a unimodal structure. We introduce IMED-UB, a algorithm that optimally exploits the unimodal-structure, by adapting to this setting the Indexed Minimum Empirical Divergence (IMED) algorithm introduced by Honda and Takemura [2015]. Owing to our proof technique, we are able to provide a concise finite-time analysis of IMED-UB algorithm. Numerical experiments show that IMED-UB competes with the state-of-the-art algorithms.

Natural language generation from structured data mainly focuses on surface-level descriptions, suffering from uncontrollable content selection and low fidelity. Previous works leverage logical forms to facilitate logical knowledge-conditioned text generation. Though achieving remarkable progress, they are data-hungry, which makes the adoption for real-world applications challenging with limited data. To this end, this paper proposes a unified framework for logical knowledge-conditioned text generation in the few-shot setting. With only a few seeds logical forms (e.g., 20/100 shot), our approach leverages self-training and samples pseudo logical forms based on content and structure consistency. Experimental results demonstrate that our approach can obtain better few-shot performance than baselines.