Datasets play a critical role in shaping the perception of performance and progress in machine learning (ML)--the way we collect, process, and analyze data affects the way we benchmark success and form new research agendas (Paullada et al., 2020; Dotan & Milli, 2020). A growing appreciation of this determinative role of datasets has sparked a concomitant concern that standard datasets used for training and evaluating ML models lack diversity along significant dimensions, for example, geography, gender, and skin type (Shankar et al., 2017; Buolamwini & Gebru, 2018). Lack of diversity in evaluation data can obfuscate disparate performance when evaluating based on aggregate accuracy (Buolamwini & Gebru, 2018). Lack of diversity in training data can limit the extent to which learned models can adequately apply to all portions of a population, a concern highlighted in recent work in the medical domain (Habib et al., 2019; Hofmanninger et al., 2020). Our work aims to develop a general unifying perspective on the way that dataset composition affects outcomes of machine learning systems.

This paper discusses the problem of weakly supervised learning of classification, in which instances are given weak labels that are produced by some label-corruption process. The goal is to derive conditions under which loss functions for weak-label learning are proper and lower-bounded -- two essential requirements for the losses used in class-probability estimation. To this end, we derive a representation theorem for proper losses in supervised learning, which dualizes the Savage representation. We use this theorem to characterize proper weak-label losses and find a condition for them to be lower-bounded. Based on these theoretical findings, we derive a novel regularization scheme called generalized logit squeezing, which makes any proper weak-label loss bounded from below, without losing properness. Furthermore, we experimentally demonstrate the effectiveness of our proposed approach, as compared to improper or unbounded losses. Those results highlight the importance of properness and lower-boundedness. The code is publicly available at https://github.com/yoshum/lower-bounded-proper-losses.

Self-supervised learning (SSL) is rapidly closing the gap with supervised methods on large computer vision benchmarks. A successful approach to SSL is to learn representations which are invariant to distortions of the input sample. However, a recurring issue with this approach is the existence of trivial constant representations. Most current methods avoid such collapsed solutions by careful implementation details. We propose an objective function that naturally avoids such collapse by measuring the cross-correlation matrix between the outputs of two identical networks fed with distorted versions of a sample, and making it as close to the identity matrix as possible. This causes the representation vectors of distorted versions of a sample to be similar, while minimizing the redundancy between the components of these vectors. The method is called Barlow Twins, owing to neuroscientist H. Barlow's redundancy-reduction principle applied to a pair of identical networks. Barlow Twins does not require large batches nor asymmetry between the network twins such as a predictor network, gradient stopping, or a moving average on the weight updates. It allows the use of very high-dimensional output vectors. Barlow Twins outperforms previous methods on ImageNet for semi-supervised classification in the low-data regime, and is on par with current state of the art for ImageNet classification with a linear classifier head, and for transfer tasks of classification and object detection.

The ability to perform arithmetic tasks is a remarkable trait of human intelligence and might form a critical component of more complex reasoning tasks. In this work, we investigate if the surface form of a number has any influence on how sequence-to-sequence language models learn simple arithmetic tasks such as addition and subtraction across a wide range of values. We find that how a number is represented in its surface form has a strong influence on the model's accuracy. In particular, the model fails to learn addition of five-digit numbers when using subwords (e.g., "32"), and it struggles to learn with character-level representations (e.g., "3 2"). By introducing position tokens (e.g., "3 10e1 2"), the model learns to accurately add and subtract numbers up to 60 digits. We conclude that modern pretrained language models can easily learn arithmetic from very few examples, as long as we use the proper surface representation. This result bolsters evidence that subword tokenizers and positional encodings are components in current transformer designs that might need improvement. Moreover, we show that regardless of the number of parameters and training examples, models cannot seem to learn addition rules that are independent of the length of the numbers seen during training. Abstraction and composition are two important themes in the study of human languages, made possible by different linguistic representations. Although treatments in different linguistic traditions vary, representations at the lexical, syntactic, and semantic levels are a common feature in nearly all theoretical studies of human language, and until relatively recently, these representations are explicitly "materialized" in language processing pipelines (for example, semantic role labeling takes as input a syntactic parse).

A woman is frustrated by the answer given by a digital assistant. The last year has seen no shortage of unprecedented circumstances. All aspects of our lives, from work to travel to shopping, have changed. During this massive disruption, we have (unfortunately) learned why ML Ops - the practice of machine learning (ML) in production and the management of an ML lifecycle, should not be an afterthought but rather a critical element of getting value from AI. Figure 1 below shows a simplified example of an AI model in action. First trained by data - past examples of the environment, the model is then put into the real world to make predictions on new inputs - which are implicitly assumed to be sufficiently similar to what the training examples were.

This paper briefly reviews the connections between meta-learning and self-supervised learning. Meta-learning can be applied to improve model generalization capability and to construct general AI algorithms. Self-supervised learning utilizes self-supervision from original data and extracts higher-level generalizable features through unsupervised pre-training or optimization of contrastive loss objectives. In self-supervised learning, data augmentation techniques are widely applied and data labels are not required since pseudo labels can be estimated from trained models on similar tasks. Meta-learning aims to adapt trained deep models to solve diverse tasks and to develop general AI algorithms. We review the associations of meta-learning with both generative and contrastive self-supervised learning models. Unlabeled data from multiple sources can be jointly considered even when data sources are vastly different. We show that an integration of meta-learning and self-supervised learning models can best contribute to the improvement of model generalization capability. Self-supervised learning guided by meta-learner and general meta-learning algorithms under self-supervision are both examples of possible combinations.

This article introduces the concept of Manifold Learning. A large number of machine learning datasets involve thousands and sometimes millions of features. These features can make training very slow. In addition, there is plenty of space in high dimensions making the high-dimensional datasets very sparse, as most of the training instances are quite likely to be far from each other. This increases the risk of overfitting since the predictions will be based on much larger extrapolations as compared to those on low-dimensional data.

We present a modelling framework for the investigation of supervised learning in non-stationary environments. Specifically, we model two example types of learning systems: prototype-based Learning Vector Quantization (LVQ) for classification and shallow, layered neural networks for regression tasks. We investigate so-called student teacher scenarios in which the systems are trained from a stream of high-dimensional, labeled data. Properties of the target task are considered to be non-stationary due to drift processes while the training is performed. Different types of concept drift are studied, which affect the density of example inputs only, the target rule itself, or both. By applying methods from statistical physics, we develop a modelling framework for the mathematical analysis of the training dynamics in non-stationary environments. Our results show that standard LVQ algorithms are already suitable for the training in non-stationary environments to a certain extent. However, the application of weight decay as an explicit mechanism of forgetting does not improve the performance under the considered drift processes. Furthermore, we investigate gradient-based training of layered neural networks with sigmoidal activation functions and compare with the use of rectified linear units (ReLU). Our findings show that the sensitivity to concept drift and the effectiveness of weight decay differs significantly between the two types of activation function.

Due to heightened concerns regarding the outbreak of COVID-19, we are adding more instructor-led online training courses as an alternative to classroom courses. This course is also offered in an online, self-paced format. This two-day course focuses on data analytics and machine learning techniques in MATLAB using functionality within Statistics and Machine Learning Toolbox and Deep Learning Toolbox . The course demonstrates the use of unsupervised learning to discover features in large data sets and supervised learning to build predictive models. Examples and exercises highlight techniques for visualization and evaluation of results.

Triplet loss is an extremely common approach to distance metric learning. Representations of images from the same class are optimized to be mapped closer together in an embedding space than representations of images from different classes. Much work on triplet losses focuses on selecting the most useful triplets of images to consider, with strategies that select dissimilar examples from the same class or similar examples from different classes. The consensus of previous research is that optimizing with the \textit{hardest} negative examples leads to bad training behavior. That's a problem -- these hardest negatives are literally the cases where the distance metric fails to capture semantic similarity. In this paper, we characterize the space of triplets and derive why hard negatives make triplet loss training fail. We offer a simple fix to the loss function and show that, with this fix, optimizing with hard negative examples becomes feasible. This leads to more generalizable features, and image retrieval results that outperform state of the art for datasets with high intra-class variance.