The current pandemic has forced people globally to remain in isolation and practice social distancing, which creates the need for a system to combat the resulting loneliness and negative emotions. In this paper we propose Nora, a virtual coaching platform designed to utilize natural language understanding in its dialogue system and suggest other recommendations based on user interactions. It is intended to provide assistance and companionship to people undergoing self-quarantine or work-from-home routines. Nora helps users gauge their well-being by detecting and recording the user's emotion, sentiment, and stress. Nora also recommends various workout, meditation, or yoga exercises to users in support of developing a healthy daily routine. In addition, we provide a social community inside Nora, where users can connect and share their experiences with others undergoing a similar isolation procedure. Nora can be accessed from anywhere via a web link and has support for both English and Mandarin.
Typical graph embeddings may not capture type-specific bipartite graph features that arise in such areas as recommender systems, data visualization, and drug discovery. Machine learning methods utilized in these applications would be better served with specialized embedding techniques. We propose two embeddings for bipartite graphs that decompose edges into sets of indirect relationships between node neighborhoods. When sampling higher-order relationships, we reinforce similarities through algebraic distance on graphs. We also introduce ensemble embeddings to combine both into a "best of both worlds" embedding. The proposed methods are evaluated on link prediction and recommendation tasks and compared with other state-of-the-art embeddings. While being all highly beneficial in applications, we demonstrate that none of the considered embeddings is clearly superior (in contrast to what is claimed in many papers), and discuss the trade offs present among them. Reproducibility: Our code, data sets, and results are all publicly available online at: http://bit.ly/fobe_hobe_code.
One of the most crucial issues in data mining is to model human behaviour in order to provide personalisation, adaptation and recommendation. This usually involves implicit or explicit knowledge, either by observing user interactions, or by asking users directly. But these sources of information are always subject to the volatility of human decisions, making utilised data uncertain to a particular extent. In this contribution, we elaborate on the impact of this human uncertainty when it comes to comparative assessments of different data mining approaches. In particular, we reveal two problems: (1) biasing effects on various metrics of model-based prediction and (2) the propagation of uncertainty and its thus induced error probabilities for algorithm rankings. For this purpose, we introduce a probabilistic view and prove the existence of those problems mathematically, as well as provide possible solution strategies. We exemplify our theory mainly in the context of recommender systems along with the metric RMSE as a prominent example of precision quality measures.
Offline evaluation of information retrieval and recommendation has traditionally focused on distilling the quality of a ranking into a scalar metric such as average precision or normalized discounted cumulative gain. We can use this metric to compare the performance of multiple systems for the same request. Although evaluation metrics provide a convenient summary of system performance, they also collapse subtle differences across users into a single number and can carry assumptions about user behavior and utility not supported across retrieval scenarios. We propose recall-paired preference (RPP), a metric-free evaluation method based on directly computing a preference between ranked lists. RPP simulates multiple user subpopulations per query and compares systems across these pseudo-populations. Our results across multiple search and recommendation tasks demonstrate that RPP substantially improves discriminative power while correlating well with existing metrics and being equally robust to incomplete data.
The increasing adoption of artificial intelligence requires accurate forecasts and means to understand the reasoning of artificial intelligence models behind such a forecast. Explainable Artificial Intelligence (XAI) aims to provide cues for why a model issued a certain prediction. Such cues are of utmost importance to decision-making since they provide insights on the features that influenced most certain forecasts and let the user decide if the forecast can be trusted. Though many techniques were developed to explain black-box models, little research was done on assessing the quality of those explanations and their influence on decision-making. We propose an ontology and knowledge graph to support collecting feedback regarding forecasts, forecast explanations, recommended decision-making options, and user actions. This way, we provide means to improve forecasting models, explanations, and recommendations of decision-making options. We tailor the knowledge graph for the domain of demand forecasting and validate it on real-world data.
Treatment policies learned via reinforcement learning (RL) from observational health data are sensitive to subtle choices in study design. We highlight a simple approach, trajectory inspection, to bring clinicians into an iterative design process for model-based RL studies. We inspect trajectories where the model recommends unexpectedly aggressive treatments or believes its recommendations would lead to much more positive outcomes. Then, we examine clinical trajectories simulated with the learned model and policy alongside the actual hospital course to uncover possible modeling issues. To demonstrate that this approach yields insights, we apply it to recent work on RL for inpatient sepsis management. We find that a design choice around maximum trajectory length leads to a model bias towards discharge, that the RL policy preference for high vasopressor doses may be linked to small sample sizes, and that the model has a clinically implausible expectation of discharge without weaning off vasopressors.
Although modern recommendation systems can exploit the structure in users' item feedback, most are powerless in the face of new users who provide no structure for them to exploit. In this paper we introduce ImplicitCE, an algorithm for recommending items to new users during their sign-up flow. ImplicitCE works by transforming users' implicit feedback towards auxiliary domain items into an embedding in the target domain item embedding space. ImplicitCE learns these embedding spaces and transformation function in an end-to-end fashion and can co-embed users and items with any differentiable similarity function. To train ImplicitCE we explore methods for maximizing the correlations between model predictions and users' affinities and introduce Sample Correlation Update, a novel and extremely simple training strategy. Finally, we show that ImplicitCE trained with Sample Correlation Update outperforms a variety of state of the art algorithms and loss functions on both a large scale Twitter dataset and the DBLP dataset.
Every fiscal quarter, companies hold earnings calls in which company executives respond to questions from analysts. After these calls, analysts often change their price target recommendations, which are used in equity research reports to help investors make decisions. In this paper, we examine analysts' decision making behavior as it pertains to the language content of earnings calls. We identify a set of 20 pragmatic features of analysts' questions which we correlate with analysts' pre-call investor recommendations. We also analyze the degree to which semantic and pragmatic features from an earnings call complement market data in predicting analysts' post-call changes in price targets. Our results show that earnings calls are moderately predictive of analysts' decisions even though these decisions are influenced by a number of other factors including private communication with company executives and market conditions. A breakdown of model errors indicates disparate performance on calls from different market sectors.
Efficient Maximum Inner Product Search (MIPS) is an important task that has a wide applicability in recommendation systems and classification with a large number of classes. Solutions based on locality-sensitive hashing (LSH) as well as tree-based solutions have been investigated in the recent literature, to perform approximate MIPS in sublinear time. In this paper, we compare these to another extremely simple approach for solving approximate MIPS, based on variants of the k-means clustering algorithm. Specifically, we propose to train a spherical k-means, after having reduced the MIPS problem to a Maximum Cosine Similarity Search (MCSS). Experiments on two standard recommendation system benchmarks as well as on large vocabulary word embeddings, show that this simple approach yields much higher speedups, for the same retrieval precision, than current state-of-the-art hashing-based and tree-based methods. This simple method also yields more robust retrievals when the query is corrupted by noise.