Factorization machine (FM) is an effective model for feature-based recommendation which utilizes inner product to capture second-order feature interactions. However, one of the major drawbacks of FM is that it couldn't capture complex high-order interaction signals. A common solution is to change the interaction function, such as stacking deep neural networks on the top of FM. In this work, we propose an alternative approach to model high-order interaction signals in the embedding level, namely Generalized Embedding Machine (GEM). The embedding used in GEM encodes not only the information from the feature itself but also the information from other correlated features. Under such situation, the embedding becomes high-order. Then we can incorporate GEM with FM and even its advanced variants to perform feature interactions. More specifically, in this paper we utilize graph convolution networks (GCN) to generate high-order embeddings. We integrate GEM with several FM-based models and conduct extensive experiments on two real-world datasets. The results demonstrate significant improvement of GEM over corresponding baselines.

Higher-order tensors arise frequently in applications such as neuroimaging, recommendation system, social network analysis, and psychological studies. We consider the problem of low-rank tensor estimation from possibly incomplete, ordinal-valued observations. Two related problems are studied, one on tensor denoising and another on tensor completion. We propose a multi-linear cumulative link model, develop a rank-constrained M-estimator, and obtain theoretical accuracy guarantees. Our mean squared error bound enjoys a faster convergence rate than previous results, and we show that the proposed estimator is minimax optimal under the class of low-rank models. Furthermore, the procedure developed serves as an efficient completion method which guarantees consistent recovery of an order-$K$ $(d,\ldots,d)$-dimensional low-rank tensor using only $\tilde{\mathcal{O}}(Kd)$ noisy, quantized observations. We demonstrate the outperformance of our approach over previous methods on the tasks of clustering and collaborative filtering.

Together with the UN Refugee Agency (UNHCR) 34 collaborators built several AI and machine learning based solutions to predict forced displacement, violent conflicts, and climate change in Somalia. In addition, an exploratory data analysis resulted in powerful insights regarding conflict types, areas, and reasons. The findings will help UNHCR to execute necessary support mechanism for people at need in a faster and more effective way. Millions of people in Somalia are forced to leave their current area of residence or community due to resource shortage and natural disasters like droughts and floods as well as violent conflicts. Our challenge partner, UNHCR, provides assistance and protection for those who are forcibly displaced inside of Somalia.

In its two decades in business, Yandex has been called the Russian Google, Amazon, and Spotify, mostly due to the Moscow-based tech giant's expansive reach into every nook -- including online search, music streaming, email, maps and navigation, video, and more. In 2011, Yandex launched a mobile taxi-hailing service called Yandex.Taxi, leading to the inevitable "Uber of Russia" proclamations. Then in 2017, Yandex.Taxi and Uber merged their operations in the region to launch a new joint venture targeting Eastern Europe. Yandex.Taxi now operates across the Commonwealth of Independent States (CIS), in addition to a handful of markets elsewhere in Europe, the Middle East, and Africa. The company has followed a trajectory similar to Uber's, insofar as it now also offers food delivery, and in 2018 it launched one of Europe's first public self-driving taxi services as part of a limited pilot.

Efficient and interpretable spatial analysis is crucial in many fields such as geology, sports, and climate science. Large-scale spatial data often contains complex higher-order correlations across features and locations. While tensor latent factor models can describe higher-order correlations, they are inherently computationally expensive to train. Furthermore, for spatial analysis, these models should not only be predictive but also be spatially coherent. However, latent factor models are sensitive to initialization and can yield inexplicable results. We develop a novel Multi-resolution Tensor Learning (MRTL) algorithm for efficiently learning interpretable spatial patterns. MRTL initializes the latent factors from an approximate full-rank tensor model for improved interpretability and progressively learns from a coarse resolution to the fine resolution for an enormous computation speedup. We also prove the theoretical convergence and computational complexity of MRTL. When applied to two real-world datasets, MRTL demonstrates 4 ~ 5 times speedup compared to a fixed resolution while yielding accurate and interpretable models.

Last week we took a look at computer vision; what it is, how it works, and some of the applications for computer vision in agtech. In case you missed last week's article, computer vision or machine vision typically refers to the use of machine learning or deep learning algorithms in image processing to allow a machine to "see" and identify objects around it. Different computer vision technologies may use a variety of camera types to act as the machine's "eyes" depending on the imaging requirements. In the case of fully autonomous vehicles, an accurate computer vision system is essential. In typical vehicles, hazard detection, navigation, and object avoidance all depend on a human operator.

Electroencephalography (EEG) recordings of rhythm perception might contain enough information to distinguish different rhythm types/genres or even identify the rhythms themselves. We apply convolutional neural networks (CNNs) to analyze and classify EEG data recorded within a rhythm perception study in Kigali, Rwanda which comprises 12 East African and 12 Western rhythmic stimuli – each presented in a loop for 32 seconds to 13 participants. We investigate the impact of the data representation and the pre-processing steps for this classification tasks and compare different network structures. Using CNNs, we are able to recognize individual rhythms from the EEG with a mean classification accuracy of 24.4% (chance level 4.17%) over all subjects by looking at less than three seconds from a single channel. Aggregating predictions for multiple channels, a mean accuracy of up to 50% can be achieved for individual subjects.

A popular trick for computing a pairwise co-occurrence matrix is the product of an incidence matrix and its transpose. We present an analog for higher order tuple co-occurrences using the face-splitting product, or alternately known as the transpose Khatri-Rao product. These higher order co-occurrences encode the commonality of tokens in the company of other tokens, and thus generalize the mutual information commonly studied. We demonstrate this tensor's use via a popular NLP model, and hypergraph models of similarity. Studying an implicit meaning of a collection of things via their relationships with other things of the same type is a popular technique.

This paper studies the optimal rate of estimation in a finite Gaussian location mixture model in high dimensions without separation conditions. We assume that the number of components $k$ is bounded and that the centers lie in a ball of bounded radius, while allowing the dimension $d$ to be as large as the sample size $n$. Extending the one-dimensional result of Heinrich and Kahn \cite{HK2015}, we show that the minimax rate of estimating the mixing distribution in Wasserstein distance is $\Theta((d/n)^{1/4} + n^{-1/(4k-2)})$, achieved by an estimator computable in time $O(nd^2+n^{5/4})$. Furthermore, we show that the mixture density can be estimated at the optimal parametric rate $\Theta(\sqrt{d/n})$ in Hellinger distance; however, no computationally efficient algorithm is known to achieve the optimal rate. Both the theoretical and methodological development rely on a careful application of the method of moments. Central to our results is the observation that the information geometry of finite Gaussian mixtures is characterized by the moment tensors of the mixing distribution, whose low-rank structure can be exploited to obtain a sharp local entropy bound.

The success of Convolutional Neural Networks (CNNs) in image classification has prompted efforts to study their use for classifying image data obtained in Particle Physics experiments. Here, we discuss our efforts to apply CNNs to 2D and 3D image data from particle physics experiments to classify signal from background. In this work we present an extensive convolutional neural architecture search, achieving high accuracy for signal/background discrimination for a HEP classification use-case based on simulated data from the Ice Cube neutrino observatory and an ATLAS-like detector. We demonstrate among other things that we can achieve the same accuracy as complex ResNet architectures with CNNs with less parameters, and present comparisons of computational requirements, training and inference times.