The spread of misinformation online is a challenge that may have an impact on society by misleading and undermining the trust of people in domains such as politics or public health. While fact-checking is one way to identify misinformation, it is a slow process and requires significant effort. Improving the efficiency of fact-checking by automating parts of the process or defining new processes to validate claims is a challenging task with a need for expertise from multiple disciplines. The aim of ROME 2019 is to bring together researchers from various fields such as Information Retrieval, Natural Language Processing, Semantic Web and Complex Networks to discuss these problems and define new directions in the area of automated fact-checking.

Latent factor models are increasingly popular for modeling multi-relational knowledge graphs. By their vectorial nature, it is not only hard to interpret why this class of models works so well, but also to understand where they fail and how they might be improved. We conduct an experimental survey of state-of-the-art models, not towards a purely comparative end, but as a means to get insight about their inductive abilities. To assess the strengths and weaknesses of each model, we create simple tasks that exhibit first, atomic properties of binary relations, and then, common inter-relational inference through synthetic genealogies. Based on these experimental results, we propose new research directions to improve on existing models.

Marzieh Saeidi, Max Bartolo, Patrick Lewis, Sameer Singh, Tim Rocktäschel, Mike Sheldon, Guillaume Bouchard, Sebastian Riedel. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018.

In statistical relational learning, knowledge graph completion deals with automatically understanding the structure of large knowledge graphs---labeled directed graphs---and predicting missing relationships---labeled edges. State-of-the-art embedding models propose different trade-offs between modeling expressiveness, and time and space complexity. We reconcile both expressiveness and complexity through the use of complex-valued embeddings and explore the link between such complex-valued embeddings and unitary diagonalization. We corroborate our approach theoretically and show that all real square matrices---thus all possible relation/adjacency matrices---are the real part of some unitarily diagonalizable matrix. This results opens the door to a lot of other applications of square matrices factorization. Our approach based on complex embeddings is arguably simple, as it only involves a Hermitian dot product, the complex counterpart of the standard dot product between real vectors, whereas other methods resort to more and more complicated composition functions to increase their expressiveness. The proposed complex embeddings are scalable to large data sets as it remains linear in both space and time, while consistently outperforming alternative approaches on standard link prediction benchmarks.

In this paper, we introduce the task of targeted aspect-based sentiment analysis. The goal is to extract fine-grained information with respect to entities mentioned in user comments. This work extends both aspect-based sentiment analysis – that assumes a single entity per document — and targeted sentiment analysis — that assumes a single sentiment towards a target entity. In particular, we identify the sentiment towards each aspect of one or more entities. As a testbed for this task, we introduce the SentiHood dataset, extracted from a question answering (QA) platform where urban neighbourhoods are discussed by users. In this context units of text often mention several aspects of one or more neighbourhoods. This is the first time that a generic social media platform,i.e. QA, is used for fine-grained opinion mining. Text coming from QA platforms are far less constrained compared to text from review specific platforms which current datasets are based on. We develop several strong baselines, relying on logistic regression and state-of-the-art recurrent neural networks

To learn text understanding models with millions of parameters one needs massive amounts of data.In this work, we argue that generating data can compensate for this need.While defining generic data generators is difficult, we propose to allow generators to be "weakly" specified in the sense that a set of parameters controls how the data is generated.Consider for example generators where the example templates, grammar, and/or vocabulary is determined by this set of parameters.Instead of manually tuning these parameters, we learn them from the limited training data at our disposal.To achieve this, we derive an efficient algorithm called GENERE that jointly estimates the parameters of the model and the undetermined generation parameters.We illustrate its benefits by learning to solve math exam questions using a highly parametrized sequence-to-sequence neural network.

In statistical relational learning, the link prediction problem is key to automatically understand the structure of large knowledge bases. As in previous studies, we propose to solve this problem through latent factorization. However, here we make use of complex valued embeddings. The composition of complex embeddings can handle a large variety of binary relations, among them symmetric and antisymmetric relations. Compared to state-of-the-art models such as Neural Tensor Network and Holographic Embeddings, our approach based on complex embeddings is arguably simpler, as it only uses the Hermitian dot product, the complex counterpart of the standard dot product between real vectors. Our approach is scalable to large datasets as it remains linear in both space and time, while consistently outperforming alternative approaches on standard link prediction benchmarks.

We introduce the four-parameter IBP compound Dirichlet process (ICDP), a stochastic process that generates sparse non-negative vectors with potentially an unbounded number of entries. If we repeatedly sample from the ICDP we can generate sparse matrices with an infinite number of columns and power-law characteristics. We apply the four-parameter ICDP to sparse nonparametric topic modelling to account for the very large number of topics present in large text corpora and the power-law distribution of the vocabulary of natural languages. The model, which we call latent IBP compound Dirichlet allocation (LIDA), allows for power-law distributions, both, in the number of topics summarising the documents and in the number of words defining each topic. It can be interpreted as a sparse variant of the hierarchical Pitman-Yor process when applied to topic modelling. We derive an efficient and simple collapsed Gibbs sampler closely related to the collapsed Gibbs sampler of latent Dirichlet allocation (LDA), making the model applicable in a wide range of domains. Our nonparametric Bayesian topic model compares favourably to the widely used hierarchical Dirichlet process and its heavy tailed version, the hierarchical Pitman-Yor process, on benchmark corpora. Experiments demonstrate that accounting for the power-distribution of real data is beneficial and that sparsity provides more interpretable results.

In many applications, multiple interlinked sources of data are available and they cannot be represented by a single adjacency matrix, to which large scale factorization method could be applied. Collective matrix factorization is a simple yet powerful approach to jointly factorize multiple matrices, each of which represents a relation between two entity types. Existing algorithms to estimate parameters of collective matrix factorization models are based on non-convex formulations of the problem; in this paper, a convex formulation of this approach is proposed. This enables the derivation of large scale algorithms to estimate the parameters, including an iterative eigenvalue thresholding algorithm. Numerical experiments illustrate the benefits of this new approach.

Language models can be formalized as loglinear regression models where the input features represent previously observed contexts up to a certain length m.The complexity of existing algorithms to learn the parameters by maximum likelihood scale linearly in nd, where n is the length of the training corpus and d is the number of observed features.We present a model that grows logarithmically in d, making it possible to efficiently leverage longer contexts.We account for the sequential structure of natural language using treestructured penalized objectives to avoid overfitting and achieve better generalization.

Collaborative tagging systems are now deployed extensively to help users share and organize resources. Tag prediction and recommendation can simplify and streamline the user experience, and by modeling user preferences, predictive accuracy can be significantly improved. However, previous methods typically model user behavior based only on a log of prior tags, neglecting other behaviors and information in social tagging systems, e.g., commenting on items and connecting with other users. On the other hand, little is known about the connection and correlations among these behaviors and contexts in social tagging systems.

Enterprise wikis commonly see low adoption rates, preventing them from reaching the critical mass that is needed to make them valuable. The high interaction costs for contributing content to these wikis is a key factor impeding wiki adoption. Much of the collaboration among knowledge workers continues to occur in email, which causes useful information to stay siloed in personal inboxes. In this demo we present Mail2Wiki, a system that enables easy contribution and initial curation of content from the personal space of email to the shared repository of a wiki.

The viability of cloud computing for information-intensive tasks arising in real-time opinion mining and sentiment analysis of large online text streams is described. We show how a smart distributed architecture enables an efficient and scalable design for opinion mining on internet-based content that answers key challenges, such as integrating heterogeneous data sources and adapting to events through dynamic system configuration. In particular, we present a novel approach of semantic complex event processing in a cloud environment capturing different levels of information, such as event data (e.g. content from various heterogeneous, distributed sources) as well as associations identified during the opinion mining and sentiment analysis process (e.g. dynamic co-reference resolution).

We propose a new variational EM algorithm for fitting factor analysis models with mixed continuous and categorical observations. The algorithm is based on a simple quadratic bound to the log-sum-exp function. In the special case of fully observed binary data, the bound we propose is significantly faster than previous variational methods. We show that EM is significantly more robust in the presence of missing data compared to treating the latent factors as parameters, which is the approach used by exponential family PCA and other related matrix-factorization methods. A further benefit of the variational approach is that it can easily be extended to the case of mixtures of factor analyzers, as we show. We present results on synthetic and real data sets demonstrating several desirable properties of our proposed method.

We propose a deterministic method to evaluate the integral of a positive function based on soft-binning functions that smoothly cut the integral into smaller integrals that are easier to approximate. In combination with mean-field approximations for each individual sub-part this leads to a tractable algorithm that alternates between the optimization of the bins and the approximation of the local integrals. We introduce suitable choices for the binning functions such that a standard mean field approximation can be extended to a split mean field approximation without the need for extra derivations. The method can be seen as a revival of the ideas underlying the mixture mean field approach. The latter can be obtained as a special case by taking soft-max functions for the binning.

Many types of regularization schemes have been employed in statistical learning, each one motivated by some assumption about the problem domain. In this paper, we present a unified asymptotic analysis of smooth regularizers, which allows us to see how the validity of these assumptions impacts the success of a particular regularizer. In addition, our analysis motivates an algorithm for optimizing regularization parameters, which in turn can be analyzed within our framework. We apply our analysis to several examples, including hybrid generative-discriminative learning and multi-task learning.

Given any generative classifier based on an inexact density model, we can define a discriminative counterpart that reduces its asymptotic error rate, while increasing the estimation variance. An optimal bias-variance balance might be found using hybrid generative-discriminative (HGD) approaches. In these paper, these methods are defined in a unified framework. This allow us to find sufficient conditions under which an improvement in generalization performances is guaranteed. Numerical experiments illustrate the well fondness of our statements.

In modern business, educational, and other settings, it is common to provide a digital network that interconnects hardware devices for shared access by the users (e.g., in an office where printers are available for use by all the office workers). In such a context, so-called "soft" failures, where a device silently starts working in degraded mode, may easily go un-noticedfor a long time, resulting in potential productivity loss. It is therefore advantageous to enable system administrators to identify soft failures at an early stage. We propose here a probabilistic method using variational inference on a factorial hidden Markov model to automatically discover soft failures, based on the analysis of simple usage information which is normally logged by the network infrastructure. We propose to mine these logs in order to discover statistically significant deviations in the usage behavior of the overall infrastructure, and we identify such deviations with soft failures, or, in any case, situations of interest to an administrator.

This paper is concerned with the selection of a generative model for supervised classification. Classical criteria for model selection assess the fit of a model rather than its ability to produce a low classification error rate. A new criterion, the Bayesian Entropy Criterion (BEC), is proposed. This criterion takes into account the decisional purpose of a model by minimizing the integrated classification entropy. It provides an interesting alternative to the cross-validated error rate which is computationally expensive. The asymptotic behavior of the BEC criterion is presented. Numerical experiments on both simulated and real data sets show that BEC performs better than the BIC criterion to select a model minimizing the classification error rate and provides analogous performance to the cross-validated error rate.

We propose a generative model that codes the geometry and appearance of generic visual object categories as a loose hierarchy of parts, with probabilistic spatial relations linking parts to subparts, soft assignment of subparts to parts, and scale invariant keypoint based local features at the lowest level of the hierarchy. The method is designed to efficiently handle categories containing hundreds of redundant local features, such as those returned by current key-point detectors. This robustness allows it to outperform constellation style models, despite their stronger spatial models. The model is initialized by robust bottom-up voting over location-scale pyramids, and optimized by expectation-maximization. Training is rapid, and objects do not need to be marked in the training images. Experiments on several popular datasets show the method's ability to capture complex natural object classes.