When developing models for regulated decision making, sensitive features like age, race and gender cannot be used and must be obscured from model developers to prevent bias. However, the remaining features still need to be tested for correlation with sensitive features, which can only be done with the knowledge of those features. We resolve this dilemma using a fully homomorphic encryption scheme, allowing model developers to train linear regression and logistic regression models and test them for possible bias without ever revealing the sensitive features in the clear. We demonstrate how it can be applied to leave-one-out regression testing, and show using the adult income data set that our method is practical to run.

Learning effective policies for sparse objectives is a key challenge in Deep Reinforcement Learning (RL). A common approach is to design task-related dense rewards to improve task learnability. While such rewards are easily interpreted, they rely on heuristics and domain expertise. Alternate approaches that train neural networks to discover dense surrogate rewards avoid heuristics, but are high-dimensional, black-box solutions offering little interpretability. In this paper, we present a method that discovers dense rewards in the form of low-dimensional symbolic trees - thus making them more tractable for analysis. The trees use simple functional operators to map an agent's observations to a scalar reward, which then supervises the policy gradient learning of a neural network policy. We test our method on continuous action spaces in Mujoco and discrete action spaces in Atari and Pygame environments. We show that the discovered dense rewards are an effective signal for an RL policy to solve the benchmark tasks. Notably, we significantly outperform a widely used, contemporary neural-network based reward-discovery algorithm in all environments considered.

This is a streamlined abridgement of my paper with Pierre Albert, published at LREC's Workshop on Legal and Ethical Issues in Human Language Technologies 2020. If you use any of this guide in your research, please do cite our paper titled "Ethically Collecting Multi-Modal Spontaneous Conversations with People that have Cognitive Impairments": Getting ethical approval to collect a crucial corpus took me over a year to complete. This was relatively fresh ground to tread, but I hope other researchers want to work on the accessibility of voice assistants for people with all varieties of cognitive impairments. This practical guide aims to help future researchers, like me, collect these valuable datasets quickly without compromising any ethical considerations or data security. Over a year ago now, I decided that I wanted to work to make voice assistants (Siri, Alexa, etc…) more accessible for people with dementia. To begin this project, I (with two of my supervisors) first detailed some of the critical challenges that need to be tackled if we are to make progress towards this goal.

Existing Graph Neural Network (GNN) methods that learn inductive unsupervised graph representations focus on learning node and edge representations by predicting observed edges in the graph. Although such approaches have shown advances in downstream node classification tasks, they are ineffective in jointly representing larger $k$-node sets, $k{>}2$. We propose MHM-GNN, an inductive unsupervised graph representation approach that combines joint $k$-node representations with energy-based models (hypergraph Markov networks) and GNNs. To address the intractability of the loss that arises from this combination, we endow our optimization with a loss upper bound using a finite-sample unbiased Markov Chain Monte Carlo estimator. Our experiments show that the unsupervised MHM-GNN representations of MHM-GNN produce better unsupervised representations than existing approaches from the literature.

Interest in artificial intelligence continues to explode across every industry, but few areas offer more opportunities for drastic improvement of human life than the application of machine learning and AI in healthcare and the medical field. Let's begin first with a definition. AI in healthcare and medicine means using data more effectively through machine learning algorithms to produce positive patient outcomes. The sheer amount of data created through IoT-enabled devices, the electronic medical record (EMR), and ever-expanding quantities of genetic data has made possible a large number of applications of artificial intelligence in healthcare. Check out the Harvard Business Review ranking of the potential value that these applications could bring to the healthcare industry. The underlying value of artificial intelligence is to enhance human decision-making and automate processes that are time- or resource-intensive for humans to perform.

Amid a pandemic implosion, these startups showcase the strength, diversity and adaptability of America's entrepreneurs--and provide hope for the country's economic future. Even in the most challenging times, the best entrepreneurs find ways to excel. The 25 small companies listed here--all of which have less than $50 million in 2019 sales and fewer than 200 employees--are successfully navigating this turbulent year, even as some of their founders cope with personal losses from Covid-19. Some make things that are increasingly critical, such as software that improves hospital operations or robots that clean schools. Others have shifted to adapt to the pandemic, such as the extended-stay hotel operator using its rooms to house displaced international students and traveling doctors, or the maker of rolling buffets that started producing plexiglass dividers. These small-business standouts showcase the strength, adaptability and diversity of America's entrepreneurs, giving us hope for the country's economic future.

Recent advances in neural forecasting have produced major improvements in accuracy for probabilistic demand prediction. In this work, we propose novel improvements to the current state of the art by incorporating changes inspired by recent advances in Transformer architectures for Natural Language Processing. We develop a novel decoder-encoder attention for context-alignment, improving forecasting accuracy by allowing the network to study its own history based on the context for which it is producing a forecast. We also present a novel positional encoding that allows the neural network to learn context-dependent seasonality functions as well as arbitrary holiday distances. Finally we show that the current state of the art MQ-Forecaster (Wen et al., 2017) models display excess variability by failing to leverage previous errors in the forecast to improve accuracy. We propose a novel decoder-self attention scheme for forecasting that produces significant improvements in the excess variation of the forecast.

While Bayesian Optimization (BO) is a very popular method for optimizing expensive black-box functions, it fails to leverage the experience of domain experts. This causes BO to waste function evaluations on bad design choices (e.g., machine learning hyperparameters) that the expert already knows to work poorly. To address this issue, we introduce Prior-guided Bayesian Optimization (PrBO). PrBO allows users to inject their knowledge into the optimization process in the form of priors about which parts of the input space will yield the best performance, rather than BO's standard priors over functions (which are much less intuitive for users). PrBO then combines these priors with BO's standard probabilistic model to form a pseudo-posterior used to select which points to evaluate next. We show that PrBO is around 12x faster than state-of-the-art methods without user priors and 10,000x faster than random search on a common suite of benchmarks, and achieves a new state-of-the-art performance on a real-world hardware design application. We also show that PrBO converges faster even if the user priors are not entirely accurate and that it robustly recovers from misleading priors.

Recommender systems constitute the core engine of most social network platforms nowadays, aiming to maximize user satisfaction along with other key business objectives. Twitter is no exception. Despite the fact that Twitter data has been extensively used to understand socioeconomic and political phenomena and user behaviour, the implicit feedback provided by users on Tweets through their engagements on the Home Timeline has only been explored to a limited extent. At the same time, there is a lack of large-scale public social network datasets that would enable the scientific community to both benchmark and build more powerful and comprehensive models that tailor content to user interests. By releasing an original dataset of 160 million Tweets along with engagement information, Twitter aims to address exactly that. During this release, special attention is drawn on maintaining compliance with existing privacy laws. Apart from user privacy, this paper touches on the key challenges faced by researchers and professionals striving to predict user engagements. It further describes the key aspects of the RecSys 2020 Challenge that was organized by ACM RecSys in partnership with Twitter using this dataset.

Deep neural networks (DNNs) perform very well when trained on a large amount of data [1], but large SSVEP datasets are not commonly available for open use. Our way to overcome this problem was to employ data augmentation and transfer learning techniques to train the DNNs, as both are known to improve the performances of DNNs on smaller datasets [2]. We started with an open SSVEP dataset [3], which we consider to be large in comparison with other open databases. The electroencephalography (EEG) signals where transformed into images, specifically spectrograms, using the shorttime Fourier transform (STFT). By doing so, we take advantage of the ability of convolutional DNN in classifying images, which is well documented [1]. The neural network used in this study [4] is a DCNN based on the original VGG [5].