Neo4j, the leader in graph technology, announced the latest version of Neo4j for Graph Data Science, a breakthrough that democratizes advanced graph-based machine learning (ML) techniques by leveraging deep learning and graph convolutional neural networks. Until now, few companies outside of Google and Facebook have had the AI foresight and resources to leverage graph embeddings. This powerful and innovative technique calculates the shape of the surrounding network for each piece of data inside of a graph, enabling far better machine learning predictions. Neo4j for Graph Data Science version 1.4 democratizes these innovations to upend the way enterprises make predictions in diverse scenarios from fraud detection to tracking customer or patient journey, to drug discovery and knowledge graph completion. Neo4j for Graph Data Science version 1.4 is the first and only graph-native machine learning functionality commercially available for enterprises.

Researchers from all over the world contribute to this repository as a prelude to the peer review process for publication in traditional journals. The articles listed below represent a small fraction of all articles appearing on the preprint server. They are listed in no particular order with a link to each paper along with a brief overview. Links to GitHub repos are provided when available. Especially relevant articles are marked with a "thumbs up" icon.

With increasing demand in machine learning and data science in businesses, for upgraded data strategizing there's a need for a better workflow to ensure robustness in data modelling. Machine learning has certain steps to be followed namely – data collection, data preprocessing(cleaning and feature engineering), model training, validation and prediction on the test data(which is previously unseen by model). Here testing data needs to go through the same preprocessing as training data. For this iterative process, pipelines are used which can automate the entire process for both training and testing data. It ensures reusability of the model by reducing the redundant part, thereby speeding up the process.

Probably one of the most common algorithms around, Linear Regression is a must know for Machine Learning Practitioners. This is usually a beginner's first exposure to a real Machine Learning algorithm, and knowing how it operates on a deeper level is crucial to gain a better understanding of it. So, briefly, let's break down the real question; What really is Linear Regression? Linear Regression is a supervised learning algorithm that aims at taking a linear approach at modelling the relation between a dependent variable and an independent variable. In other words, It aims to fit a linear trendline that best captures the relationship of the data, and, from this line, it can predict what the target values may be.

Distribution regression refers to the supervised learning problem where labels are only available for groups of inputs instead of individual inputs. In this paper, we develop a rigorous mathematical framework for distribution regression where inputs are complex data streams. Leveraging properties of the expected signature and a recent signature kernel trick for sequential data from stochastic analysis, we introduce two new learning techniques, one feature-based and the other kernel-based. Each is suited to a different data regime in terms of the number of data streams and the dimensionality of the individual streams. We provide theoretical results on the universality of both approaches and demonstrate empirically their robustness to irregularly sampled multivariate time-series, achieving state-of-the-art performance on both synthetic and real-world examples from thermodynamics, mathematical finance and agricultural science.

In the contextual linear bandit setting, algorithms built on the optimism principle fail to exploit the structure of the problem and have been shown to be asymptotically suboptimal. In this paper, we follow recent approaches of deriving asymptotically optimal algorithms from problem-dependent regret lower bounds and we introduce a novel algorithm improving over the state-of-the-art along multiple dimensions. We build on a reformulation of the lower bound, where context distribution and exploration policy are decoupled, and we obtain an algorithm robust to unbalanced context distributions. Then, using an incremental primal-dual approach to solve the Lagrangian relaxation of the lower bound, we obtain a scalable and computationally efficient algorithm. Finally, we remove forced exploration and build on confidence intervals of the optimization problem to encourage a minimum level of exploration that is better adapted to the problem structure. We demonstrate the asymptotic optimality of our algorithm, while providing both problem-dependent and worst-case finite-time regret guarantees. Our bounds scale with the logarithm of the number of arms, thus avoiding the linear dependence common in all related prior works. Notably, we establish minimax optimality for any learning horizon in the special case of non-contextual linear bandits. Finally, we verify that our algorithm obtains better empirical performance than state-of-the-art baselines.

The current practice in machine learning is traditionally model-centric, casting problems as optimization over model parameters, all the while assuming the data is either fixed, or subject to extrinsic and inevitable change. On one hand, this paradigm fails to capture important existing aspects of machine learning, such as the substantial data manipulation (\emph{e.g.}, augmentation) that goes into most state-of-the-art pipelines. On the other hand, this viewpoint is ill-suited to formalize novel data-centric problems, such as model-agnostic transfer learning or dataset synthesis. In this work, we view these and other problems through the lens of \textit{dataset optimization}, casting them as optimization over data-generating distributions. We approach this class of problems through Wasserstein gradient flows in probability space, and derive practical and efficient particle-based methods for a flexible but well-behaved class of objective functions. Through various experiments on synthetic and real datasets, we show that this framework provides a principled and effective approach to dataset shaping, transfer, and interpolation.

Flexible variational distributions improve variational inference but are harder to optimize. In this work we present a control variate that is applicable for any reparameterizable distribution with known mean and covariance matrix, e.g. Gaussians with any covariance structure. The control variate is based on a quadratic approximation of the model, and its parameters are set using a double-descent scheme by minimizing the gradient estimator's variance. We empirically show that this control variate leads to large improvements in gradient variance and optimization convergence for inference with non-factorized variational distributions.

Partial-label learning (PLL) is a multi-class classification problem, where each training example is associated with a set of candidate labels. Even though many practical PLL methods have been proposed in the last two decades, there lacks a theoretical understanding of the consistency of those methods--none of the PLL methods hitherto possesses a generation process of candidate label sets, and then it is still unclear why such a method works on a specific dataset and when it may fail given a different dataset. In this paper, we propose the first generation model of candidate label sets, and develop two novel PLL methods that are guaranteed to be provably consistent, i.e., one is risk-consistent and the other is classifier-consistent. Our methods are advantageous, since they are compatible with any deep network or stochastic optimizer. Furthermore, thanks to the generation model, we would be able to answer the two questions above by testing if the generation model matches given candidate label sets. Experiments on benchmark and real-world datasets validate the effectiveness of the proposed generation model and two PLL methods.

Varying domains and biased datasets can lead to differences between the training and the target distributions, known as covariate shift. Current approaches for alleviating this often rely on estimating the ratio of training and target probability density functions. These techniques require parameter tuning and can be unstable across different datasets. We present a new method for handling covariate shift using the empirical cumulative distribution function estimates of the target distribution by a rigorous generalization of a recent idea proposed by Vapnik and Izmailov. Further, we show experimentally that our method is more robust in its predictions, is not reliant on parameter tuning and shows similar classification performance compared to the current state-of-the-art techniques on synthetic and real datasets.