For machine learning systems to be reliable, we must understand their performance in unseen, out-of-distribution environments. In this paper, we empirically show that out-of-distribution performance is strongly correlated with in-distribution performance for a wide range of models and distribution shifts. Specifically, we demonstrate strong correlations between in-distribution and out-of-distribution performance on variants of CIFAR-10 & ImageNet, a synthetic pose estimation task derived from YCB objects, satellite imagery classification in FMoW-WILDS, and wildlife classification in iWildCam-WILDS. The strong correlations hold across model architectures, hyperparameters, training set size, and training duration, and are more precise than what is expected from existing domain adaptation theory. To complete the picture, we also investigate cases where the correlation is weaker, for instance some synthetic distribution shifts from CIFAR-10-C and the tissue classification dataset Camelyon17-WILDS. Finally, we provide a candidate theory based on a Gaussian data model that shows how changes in the data covariance arising from distribution shift can affect the observed correlations.

Continual learning-the ability to learn many tasks in sequence-is critical for artificial learning systems. Yet standard training methods for deep networks often suffer from catastrophic forgetting, where learning new tasks erases knowledge of earlier tasks. While catastrophic forgetting labels the problem, the theoretical reasons for interference between tasks remain unclear. Here, we attempt to narrow this gap between theory and practice by studying continual learning in the teacher-student setup. We extend previous analytical work on two-layer networks in the teacher-student setup to multiple teachers. Using each teacher to represent a different task, we investigate how the relationship between teachers affects the amount of forgetting and transfer exhibited by the student when the task switches. In line with recent work, we find that when tasks depend on similar features, intermediate task similarity leads to greatest forgetting. However, feature similarity is only one way in which tasks may be related. The teacher-student approach allows us to disentangle task similarity at the level of readouts (hidden-to-output weights) and features (input-to-hidden weights). We find a complex interplay between both types of similarity, initial transfer/forgetting rates, maximum transfer/forgetting, and long-term transfer/forgetting. Together, these results help illuminate the diverse factors contributing to catastrophic forgetting.

The Fisher information is one of the most fundamental concepts in statistical machine learning. Intuitively, it measures the amount of information carried by a single random observation when the underlying model varies along certain directions in the parameter space: if such a variation does not change the underlying model, then a corresponding observation contains zero (Fisher) information and is non-informative regarding the varied parameter. Parameter estimation is impossible in this case. Otherwise, if the variation significantly changes the model and has large information, then an observation is informative and the parameter estimation can be more efficient as compared to parameters with small Fisher information. In machine learning, this basic concept is useful for defining intrinsic structures of the parameter space, measuring model complexity, and performing gradient-based optimization.

"Uncle Stock is one of the best stock screeners up there and I have tested a few. Advising what to do next is a must for my stock market investment. "I would like to thank you immensely for creating such a marvellous gem which contains extensive details preventing us from undergoing the pain to calculate them. This is simply mind blowing. I have everything that I would ever need with tremendous customisation options.

A Chinese gene company selling prenatal tests around the world developed them in collaboration with the country's military and is using them to collect genetic data from millions of women for sweeping research on the traits of populations, a review of scientific papers and company statements found. U.S. government advisers warned in March that a vast bank of genomic data that the company, BGI Group, is amassing and analyzing with artificial intelligence could give China a path to economic and military advantage. As science pinpoints new links between genes and human traits, access to the biggest, most diverse set of human genomes is a strategic edge. The technology could propel China to dominate global pharmaceuticals, and also potentially lead to genetically enhanced soldiers, or engineered pathogens to target the U.S. population or food supply, the advisers said. Reuters has found that BGI's prenatal test, one of the most popular in the world, is a source of genetic data for the company, which has worked with the Chinese military to improve "population quality" and on genetic research to combat hearing loss and altitude sickness in soldiers. BGI says it stores and reanalyzes left-over blood samples and genetic data from the prenatal tests, sold in at least 52 countries to detect abnormalities such as Down's syndrome in the fetus. The tests -- branded NIFTY for "non-invasive fetal trisomY" -- also capture genetic information about the mother, as well as personal details such as her country, height and weight, but not her name, BGI computer code shows.

And I am no committed die-hard Marxist either. In this paper I am merely asking if our planning, evaluations & reviews of investments made in education, employment and human capital from tax payers' money over the years till now (especially since 1991) been judicious enough to warrant comfort in future outputs. Inviting my readers to do a self (mental) due diligence of achievements and the progress made in our country in the past few decades as I do, all I am asking is if, given the commitments radiating among our warring political parties under an archaic political system, the future of our grandchildren safe enough? Or, given they will not join the emerging lumpen elements, ought we to plan their migration to as bizarre countries as Taiwan, China, South Korea?] "Bureaucracy served Man well in the past. But the nature of Work has changed and management must change for us to survive. Our goal is to move from a bureaucratic model that is focused on maximizing compliance to one that is focused on maximizing contribution"– Management Guru Gary Hamel, speaking on Humanocracy at an Open Interactive pop up on 18th February 2021.

A new deep learning model could help radiologists in any facility interpret chest X-rays. In a new study published in The Lancet Digital Health, investigators from Australia outlined their new tool. It is designed to alleviate heavy workloads and make it easier for providers who do not have specialty thoracic training to read these scans while reducing errors. Chest X-rays are already the most common imaging study worldwide, and that number is growing, said the team from annalise.ai, the company that created the AI model. Developing a tool to help shoulder the weight and process the workload will be critical. "The ability of the AI model to identify findings on chest X-rays is very encouraging," said Catherine Jones, MBBS, thoracic radiologist, chest lead at annalise.ai,

Coupled with biaffine decoders, transformers have been effectively adapted to text-to-graph transduction and achieved state-of-the-art performance on AMR parsing. Many prior works, however, rely on the biaffine decoder for either or both arc and label predictions although most features used by the decoder may be learned by the transformer already. This paper presents a novel approach to AMR parsing by combining heterogeneous data (tokens, concepts, labels) as one input to a transformer to learn attention, and use only attention matrices from the transformer to predict all elements in AMR graphs (concepts, arcs, labels). Although our models use significantly fewer parameters than the previous state-of-the-art graph parser, they show similar or better accuracy on AMR 2.0 and 3.0.

Although many techniques have been applied to matrix factorization (MF), they may not fully exploit the feature structure. In this paper, we incorporate the grouping effect into MF and propose a novel method called Robust Matrix Factorization with Grouping effect (GRMF). The grouping effect is a generalization of the sparsity effect, which conducts denoising by clustering similar values around multiple centers instead of just around 0. Compared with existing algorithms, the proposed GRMF can automatically learn the grouping structure and sparsity in MF without prior knowledge, by introducing a naturally adjustable non-convex regularization to achieve simultaneous sparsity and grouping effect. Specifically, GRMF uses an efficient alternating minimization framework to perform MF, in which the original non-convex problem is first converted into a convex problem through Difference-of-Convex (DC) programming, and then solved by Alternating Direction Method of Multipliers (ADMM). In addition, GRMF can be easily extended to the Non-negative Matrix Factorization (NMF) settings. Extensive experiments have been conducted using real-world data sets with outliers and contaminated noise, where the experimental results show that GRMF has promoted performance and robustness, compared to five benchmark algorithms.

Importantly, our approach does not make Here, we propose a general method for probabilistic any a-priori assumptions on the underlying distribution of time series forecasting. We combine an our data. The probabilistic output of our model is generated autoregressive recurrent neural network to model via Implicit Quantile Networks (Dabney et al., 2018) temporal dynamics with Implicit Quantile Networks (IQN) and is trained by minimizing the integrand of the to learn a large class of distributions over a Continuous Ranked Probability Score (CRPS) (Matheson & time-series target. When compared to other probabilistic Winkler, 1976).