We introduce a novel random projection technique for efficiently reducing the dimension of very high-dimensional tensors. Building upon classical results on Gaussian random projections and Johnson-Lindenstrauss transforms~(JLT), we propose two tensorized random projection maps relying on the tensor train~(TT) and CP decomposition format, respectively. The two maps offer very low memory requirements and can be applied efficiently when the inputs are low rank tensors given in the CP or TT format. Our theoretical analysis shows that the dense Gaussian matrix in JLT can be replaced by a low-rank tensor implicitly represented in compressed form with random factors, while still approximately preserving the Euclidean distance of the projected inputs. In addition, our results reveal that the TT format is substantially superior to CP in terms of the size of the random projection needed to achieve the same distortion ratio. Experiments on synthetic data validate our theoretical analysis and demonstrate the superiority of the TT decomposition.

We consider the task of auditing ML models for individual bias/unfairness. We formalize the task in an optimization problem and develop a suite of inferential tools for the optimal value. Our tools permit us to obtain asymptotic confidence intervals and hypothesis tests that cover the target/control the Type I error rate exactly. To demonstrate the utility of our tools, we use them to reveal the gender and racial biases in Northpointe's COMPAS recidivism prediction instrument.

We look at a dataset of tropical storm data in the Atlantic Ocean from 1982 to 2017 and perform deep learning predictions with uncertainty bounds on trajectories of the storms. The result of these storms, particularly the strongest ones called hurricanes--defined as having wind speeds exceeding 74 mph--can be devastating because of their strong winds and heavy precipitation that can cause dangerous tides. Tropical storms can cause major environmental disasters when they reach land, such as the 2005 Hurricane Katrina that resulted in over 850 deaths and caused major economic damage and the 2012 Hurricane Sandy that caused almost $70 billion in damage across much of the eastern United States, with peak winds of 115 mph (Hurricane). According to the National Oceanic and Atmospheric Administration, it is likely that global warming will cause hurricanes in the upcoming century to be more intense by 1 to 10% globally (with higher peak winds and lower central pressures), which will result in a higher proportion of more severe storms (NOAA). Historically, hurricane trajectory predictions have used statistical methods that can be limiting because of the nonlinearity and complexity of atmospheric systems. Deep learning techniques and specifically recurrent neural networks have grown in popularity in recent years as a strong method for approaching prediction problems because of the ability to extract important features and relationships from complex high-dimensional data, especially for forecasting and classification (McDermott and Wikle, 2019). We implemented a number of improvements over previous deep learning prediction work (Alemany et al., 2019), including predicting exact storm locations in latitude/longitude instead of a grid value and using a prediction window that uses all previous hurricane data rather than a fixed-size sliding window. While hurricane trajectory predictions have seen improvements recently (SHIPS), we build on previous work to include a fundamental uncertainty measure in the prediction for the first time as part of a neural network framework. The uncertainty measure is especially valuable for understanding a defined location range rather than only a point estimate, which is important for evacuation and safety/preparation purposes.

As machine learning becomes democratized in the era of Software 2.0, one of the most serious bottlenecks is collecting enough labeled data to ensure accurate and fair models. Recent techniques including crowdsourcing provide cost-effective ways to gather such data. However, simply collecting data as much as possible is not necessarily an effective strategy for optimizing accuracy and fairness. For example, if an online app store has enough training data for certain slices of data (say American customers), but not for others, collecting more American customer data will only bias the model training. Instead, we contend that one needs to selectively collect data and propose Slice Tuner, which collects possibly-different amounts of data per slice such that the model accuracy and fairness on all slices are optimized. At its core, Slice Tuner maintains learning curves of slices that estimate the model accuracies given more data and uses convex optimization to find the best data collection strategy. The key challenges of estimating learning curves are that they may be inaccurate if there is not enough data, and there may be dependencies among slices where collecting data for one slice influences the learning curves of others. We solve these issues by iteratively and efficiently updating the learning curves as more data is collected. We evaluate Slice Tuner on real datasets using crowdsourcing for data collection and show that Slice Tuner significantly outperforms baselines in terms of model accuracy and fairness, even for initially small slices. We believe Slice Tuner is a practical tool for suggesting concrete action items based on model analysis.

Gradient boosted trees and other regression tree models perform well in a wide range of real-world, industrial applications. These tree models (i) offer insight into important prediction features, (ii) effectively manage sparse data, and (iii) have excellent prediction capabilities. Despite their advantages, they are generally unpopular for decision-making tasks and black-box optimization, which is due to their difficult-to-optimize structure and the lack of a reliable uncertainty measure. ENTMOOT is our new framework for integrating (already trained) tree models into larger optimization problems. The contributions of ENTMOOT include: (i) explicitly introducing a reliable uncertainty measure that is compatible with tree models, (ii) solving the larger optimization problems that incorporate these uncertainty aware tree models, (iii) proving that the solutions are globally optimal, i.e. no better solution exists. In particular, we show how the ENTMOOT approach allows a simple integration of tree models into decision-making and black-box optimization, where it proves as a strong competitor to commonly-used frameworks.

Reinforcement learning has been at the center of some of the biggest artificial intelligence(AI) breakthroughs of the last five years. In mastering games like Go, Quake III or StarCraft, reinforcement learning models demonstrated that they can surpass human performance and create unique long-term strategies never explored before. Part of the magic of reinforcement learning relies on regularly rewarding the agents for actions that lead to a better outcome. That models works great in dense reward environments like games in which almost every action correspond to a specific feedback but what happens if that feedback is not available? In reinforcement learning this is known as sparse rewards environments and, unfortunately, it's a representation of most real-world scenarios.

As night falls in London, Georgina Rowlands and Anna Hart start applying makeup. Rowlands has long narrow blue triangles and thin white rectangles criss-crossing her face. Hart has a collection of red, orange and white angular shapes on hers. They're two of the four founders of the Dazzle Club, a group of artists set up last year to provoke discussion about the growing using of facial recognition technology. The group holds monthly silent walks through different parts of London to raise awareness about the technology, which they say is being used for "rampant surveillance."

Rediet Abebe uses algorithms and AI to improve access to opportunity for historically marginalized communities. When Abebe moved from her native Ethiopia to the United States to attend Harvard College, she was struck by how vital resources often fail to reach the most vulnerable people, even in the world's wealthiest nation. She now uses computational techniques to mitigate socioeconomic inequalities. While she was an intern at Microsoft, Abebe formulated an AI project that analyzes search queries to shed light on the unmet health information needs of people in Africa. Her study revealed such information as which demographic groups are likely to show interest in natural cures for HIV and which countries' residents are especially concerned about HIV/AIDS stigma and discrimination.

Social media channels, such as Facebook, Twitter, and Instagram, have altered our world forever. People are now increasingly connected than ever and reveal a sort of digital persona. Although social media certainly has several remarkable features, the demerits are undeniable as well. Recent studies have indicated a correlation between high usage of social media sites and increased depression. The present study aims to exploit machine learning techniques for detecting a probable depressed Twitter user based on both, his/her network behavior and tweets. For this purpose, we trained and tested classifiers to distinguish whether a user is depressed or not using features extracted from his/ her activities in the network and tweets. The results showed that the more features are used, the higher are the accuracy and F-measure scores in detecting depressed users. This method is a data-driven, predictive approach for early detection of depression or other mental illnesses. This study's main contribution is the exploration part of the features and its impact on detecting the depression level.

Bias in Word Embeddings has been a subject of recent interest, along with efforts for its reduction. Current approaches show promising progress towards debiasing single bias dimensions such as gender or race. In this paper, we present a joint multiclass debiasing approach that is capable of debiasing multiple bias dimensions simultaneously. In that direction, we present two approaches, HardWEAT and SoftWEAT, that aim to reduce biases by minimizing the scores of the Word Embeddings Association Test (WEAT). We demonstrate the viability of our methods by debiasing Word Embeddings on three classes of biases (religion, gender and race) in three different publicly available word embeddings and show that our concepts can both reduce or even completely eliminate bias, while maintaining meaningful relationships between vectors in word embeddings. Our work strengthens the foundation for more unbiased neural representations of textual data.