In attempts to produce machine learning models less reliant on spurious patterns in training data, researchers have recently proposed a human-in-the-loop process for generating counterfactually augmented datasets. As applied in NLP, given some documents and their (initial) labels, humans are tasked with revising the text to make a (given) counterfactual label applicable. Importantly, the instructions prohibit edits that are not necessary to flip the applicable label. Models trained on the augmented (original and revised) data have been shown to rely less on semantically irrelevant words and to generalize better out-of-domain. While this work draws on causal thinking, casting edits as interventions and relying on human understanding to assess outcomes, the underlying causal model is not clear nor are the principles underlying the observed improvements in out-of-domain evaluation. In this paper, we explore a toy analog, using linear Gaussian models. Our analysis reveals interesting relationships between causal models, measurement noise, out-of-domain generalization, and reliance on spurious signals. Interestingly our analysis suggests that data corrupted by adding noise to causal features will degrade out-of-domain performance, while noise added to non-causal features may make models more robust out-of-domain. This analysis yields interesting insights that help to explain the efficacy of counterfactually augmented data. Finally, we present a large-scale empirical study that supports this hypothesis.

Most of existing deep learning models rely on excessive amounts of labeled training data in order to achieve state-of-the-art results, even though these data can be hard or costly to get in practice. One attractive alternative is to learn with little supervision, commonly referred to as few-shot learning (FSL), and, in particular, meta-learning that learns to learn with few data from related tasks. Despite the practical success of meta-learning, many of its algorithmic solutions proposed in the literature are based on sound intuitions, but lack a solid theoretical analysis of the expected performance on the test task. In this paper, we review the recent advances in meta-learning theory and show how they can be used in practice both to better understand the behavior of popular meta-learning algorithms and to improve their generalization capacity. This latter is achieved by integrating the theoretical assumptions ensuring efficient meta-learning in the form of regularization terms into several popular meta-learning algorithms for which we provide a large study of their behavior on classic few-shot classification benchmarks. To the best of our knowledge, this is the first contribution that puts the most recent learning bounds of meta-learning theory into practice for the popular task of few-shot classification.

We introduce and study semantic capacity of terms. For example, the semantic capacity of artificial intelligence is higher than that of linear regression since artificial intelligence possesses a broader meaning scope. Understanding semantic capacity of terms will help many downstream tasks in natural language processing. For this purpose, we propose a two-step model to investigate semantic capacity of terms, which takes a large text corpus as input and can evaluate semantic capacity of terms if the text corpus can provide enough co-occurrence information of terms. Extensive experiments in three fields demonstrate the effectiveness and rationality of our model compared with well-designed baselines and human-level evaluations.

Note to the reader: This article assumes some familiarity with data science and analytics concepts. The finance team of a major Australian retailer wanted to know the answer to the question of "How much does the weather affects the weekly sales?" However, nobody in the business has a scientific method of solving it. As pointed above the business value of solving this problem is tremendous, as this could help better understand the impact of certain weather events on the top and bottom line. Not just that, the business can better monitor and report weekly transactional habits of the customer, which are influenced by the weather.

These effects are often shown using the test set, something which is considered (very) bad practice but helps for educational purposes. Welcome to another installment of the'Exploring NASA's turbofan dataset' series. This will be the third analysis on FD001, where all engines run on the same operating condition and develop the same fault. Initially we assumed the Remaining Useful Life (RUL) of the engines to decline linearly. Clipping the RUL improved the baseline linear regression by 31% (from an RMSE of 31.95 to an RMSE of 21.90). We then switched to a Support Vector Regression and squeezed out another 6% improvement for a total RMSE of 20.54.

Data science is an interdisciplinary field that combines methods from statistics, mathematics, computer science, and information science to extract insights and knowledge from data. As big data continues to surge across industries, being a data scientist is one of the most highly demanded jobs on the market. In this post, we will discuss two free online resources that are great for data science beginners. This is an introductory text available for free online as an eBook. It is written by statistics professors at Stanford University, the University of Washington, and the University of Southern California.

Over-parametrization has become a popular technique in deep learning. It is observed that by over-parametrization, a larger neural network needs a fewer training iterations than a smaller one to achieve a certain level of performance -- namely, over-parametrization leads to acceleration in optimization. However, despite that over-parametrization is widely used nowadays, little theory is available to explain the acceleration due to over-parametrization. In this paper, we propose understanding it by studying a simple problem first. Specifically, we consider the setting that there is a single teacher neuron with quadratic activation, where over-parametrization is realized by having multiple student neurons learn the data generated from the teacher neuron. We provably show that over-parametrization helps the iterate generated by gradient descent to enter the neighborhood of a global optimal solution that achieves zero testing error faster. On the other hand, we also point out an issue regarding the necessity of over-parametrization and study how the scaling of the output neurons affects the convergence time.

Incorporating a so-called "momentum" dynamic in gradient descent methods is widely used in neural net training as it has been broadly observed that, at least empirically, it often leads to significantly faster convergence. At the same time, there are very few theoretical guarantees in the literature to explain this apparent acceleration effect. In this paper we show that Polyak's momentum, in combination with over-parameterization of the model, helps achieve faster convergence in training a one-layer ReLU network on $n$ examples. We show specifically that gradient descent with Polyak's momentum decreases the initial training error at a rate much faster than that of vanilla gradient descent. We provide a bound for a fixed sample size $n$, and we show that gradient descent with Polyak's momentum converges at an accelerated rate to a small error that is controllable by the number of neurons $m$. Prior work [DZPS19] showed that using vanilla gradient descent, and with a similar method of over-parameterization, the error decays as $(1-\kappa_n)^t$ after $t$ iterations, where $\kappa_n$ is a problem-specific parameter. Our result shows that with the appropriate choice of parameters one has a rate of $(1-\sqrt{\kappa_n})^t$. This work establishes that momentum does indeed speed up neural net training.

There has been an increasingly popular trend in Universities for curriculum transformation to make teaching more interactive and suitable for online courses. An increase in the popularity of online courses would result in an increase in the number of course-related queries for academics. This, coupled with the fact that if lectures were delivered in a video on demand format, there would be no fixed time where the majority of students could ask questions. When questions are asked in a lecture there is a negligible chance of having similar questions repeatedly, but asynchronously this is more likely. In order to reduce the time spent on answering each individual question, clustering them is an ideal choice. There are different unsupervised models fit for text clustering, of which the Latent Dirichlet Allocation model is the most commonly used. We use the Hierarchical Dirichlet Process to determine an optimal topic number input for our LDA model runs. Due to the probabilistic nature of these topic models, the outputs of them vary for different runs. The general trend we found is that not all the topics were being used for clustering on the first run of the LDA model, which results in a less effective clustering. To tackle probabilistic output, we recursively use the LDA model on the effective topics being used until we obtain an efficiency ratio of 1. Through our experimental results we also establish a reasoning on how Zeno's paradox is avoided.

As Machine Learning technologies become increasingly used in contexts that affect citizens, companies as well as researchers need to be confident that their application of these methods will not have unexpected social implications, such as bias towards gender, ethnicity, and/or people with disabilities. There is significant literature on approaches to mitigate bias and promote fairness, yet the area is complex and hard to penetrate for newcomers to the domain. This article seeks to provide an overview of the different schools of thought and approaches to mitigating (social) biases and increase fairness in the Machine Learning literature. It organises approaches into the widely accepted framework of pre-processing, in-processing, and post-processing methods, subcategorizing into a further 11 method areas. Although much of the literature emphasizes binary classification, a discussion of fairness in regression, recommender systems, unsupervised learning, and natural language processing is also provided along with a selection of currently available open source libraries. The article concludes by summarising open challenges articulated as four dilemmas for fairness research.