A collaborative filtering system (CF) aims at filtering huge amount of information, in order to guide users of web applications towards items that might interest them. Such a system, consists in ...recommending a set of personalized items for an active user, according to the preferences of other similar users. Existing methods, such as memory and Matrix Factorization (MF) approaches can achieve very good recommendation accuracy, unfortunately they are computationally very expensive. Applying such approaches to real-world applications in which users, items and ratings are frequently updated remains therefore a challenge. To address this problem, we propose a novel efficient incremental CF system, based on a weighted clustering approach. Our system is designed to provide a high quality of recommendations with a very low computation cost. In contrast to existing incremental methods, the complexity of our approach does not depend on the number of users and items. Our CF system is therefore suitable for dynamic settings, involving huge databases, in which available information evolves rapidly (i.e, submission of new ratings, update of existing ratings, appearance of new users and new items). Numerical experiments, conducted on several real-world datasets, confirm the efficiency and the effectiveness of our method, by demonstrating that it is significantly better than existing incremental CF methods in terms of both scalability and recommendation quality.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPUK, ZRSKP
2.
Directional co-clustering Salah, Aghiles; Nadif, Mohamed
Advances in data analysis and classification,
09/2019, Volume:
13, Issue:
3
Journal Article
Peer reviewed
Co-clustering addresses the problem of simultaneous clustering of both dimensions of a data matrix. When dealing with high dimensional sparse data, co-clustering turns out to be more beneficial than ...one-sided clustering even if one is interested in clustering along one dimension only. Aside from being high dimensional and sparse, some datasets, such as document-term matrices, exhibit directional characteristics, and the
L
2
normalization of such data, so that it lies on the surface of a unit hypersphere, is useful. Popular co-clustering assumptions such as Gaussian or Multinomial are inadequate for this type of data. In this paper, we extend the scope of co-clustering to directional data. We present Diagonal Block Mixture of Von Mises–Fisher distributions (dbmovMFs), a co-clustering model which is well suited for directional data lying on a unit hypersphere. By setting the estimate of the model parameters under the maximum likelihood (ML) and classification ML approaches, we develop a class of EM algorithms for estimating dbmovMFs from data. Extensive experiments, on several real-world datasets, confirm the advantage of our approach and demonstrate the effectiveness of our algorithms.
Full text
Available for:
EMUNI, FIS, FZAB, GEOZS, GIS, IJS, IMTLJ, KILJ, KISLJ, MFDPS, NLZOH, NUK, OBVAL, OILJ, PNG, SAZU, SBCE, SBJE, SBMB, SBNM, UKNU, UL, UM, UPUK, VKSCE, ZAGLJ
Collaborative filtering (CF) is a widely used technique to guide the users of web applications towards items that might interest them. CF approaches are severely challenged by the characteristics of ...user-item preference matrices, which are often high dimensional and extremely sparse. Recently, several works have shown that incorporating information from social networks—such as friendship and trust relationships—into traditional CF alleviates the sparsity related issues and yields a better recommendation quality, in most cases. More interestingly, even with comparable performances, social-based CF is more beneficial than traditional CF; the former makes it possible to provide recommendations for cold start users. In this paper, we propose a novel model that leverages information from social networks to improve recommendations. While existing social CF models are based on popular modelling assumptions such as Gaussian or Multinomial, our model builds on the von Mises–Fisher assumption which turns out to be more adequate, than the aforementioned assumptions, for high dimensional sparse data. Setting the estimate of the model parameters under the maximum likelihood approach, we derive a scalable learning algorithm for analyzing data with our model. Empirical results on several real-world datasets provide strong support for the advantages of the proposed model.
Non-negative Matrix Factorization (NMF) and its variants have been successfully used for clustering text documents. However, NMF approaches like other models do not explicitly account for the ...contextual dependencies between words. To remedy this limitation, we draw inspiration from neural word embedding and posit that words that frequently co-occur within the same context (e.g., sentence or document) are likely related to each other in some semantic aspect. We then propose to jointly factorize the document-word and word-word co-occurrence matrices. The decomposition of the latter matrix encourages frequently co-occurring words to have similar latent representations and thereby reflecting the relationships among them. Empirical results, on several real-world datasets, provide strong support for the benefits of our approach. Our main finding is that we can drastically improve the clustering performance of NMF by leveraging the contextual relationships among words explicitly.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPUK, ZAGLJ, ZRSKP
Multimodal recommender systems alleviate the sparsity of historical user–item interactions. They are commonly catalogued based on the type of auxiliary data (modality) they leverage, such as ...preference data plus user-network (social), user/item texts (textual), or item images (visual), respectively. One consequence of this categorization is the tendency for virtual walls to arise between modalities. For instance, a study involving images would compare to only baselines ostensibly designed for images. However, a closer look at existing models’ statistical assumptions about any one modality would reveal that many could work just as well with other modalities. Therefore, we pursue a systematic investigation into several research questions: which modality one should rely on, whether a model designed for one modality may work with another, which model to use for a given modality. We conduct cross-modality and cross-model comparisons and analyses, yielding insightful results pointing to interesting future research directions for multimodal recommender systems.
To improve users' experience as they navigate the myriad of options offered by online marketplaces, it is essential to have well-organized product catalogs. One key ingredient to that is the ...availability of product attributes such as color or material. However, on some marketplaces such as Rakuten-Ichiba, which we focus on, attribute information is often incomplete or even missing. One promising solution to this problem is to rely on deep models pre-trained on large corpora to predict attributes from unstructured data, such as product descriptive texts and images (referred to as modalities in this paper). However, we find that achieving satisfactory performance with this approach is not straightforward but rather the result of several refinements, which we discuss in this paper. We provide a detailed description of our approach to attribute extraction, from investigating strong single-modality methods, to building a solid multimodal model combining textual and visual information. One key component of our multimodal architecture is a novel approach to seamlessly combine modalities, which is inspired by our single-modality investigations. In practice, we notice that this new modality-merging method may suffer from a modality collapse issue, i.e., it neglects one modality. Hence, we further propose a mitigation to this problem based on a principled regularization scheme. Experiments on Rakuten-Ichiba data provide empirical evidence for the benefits of our approach, which has been also successfully deployed to Rakuten-Ichiba. We also report results on publicly available datasets showing that our model is competitive compared to several recent multimodal and unimodal baselines.
La classification automatique, qui consiste à regrouper des objets similaires au sein de groupes, également appelés classes ou clusters, est sans aucun doute l’une des méthodes d’apprentissage ...non-supervisé les plus utiles dans le contexte du Big Data. En effet, avec l’expansion des volumes de données disponibles, notamment sur le web, la classification ne cesse de gagner en importance dans le domaine de la science des données pour la réalisation de différentes tâches, telles que le résumé automatique, la réduction de dimension, la visualisation, la détection d’anomalies, l’accélération des moteurs de recherche, l’organisation d’énormes ensembles de données, etc. De nombreuses méthodes de classification ont été développées à ce jour, ces dernières sont cependant fortement mises en difficulté par les caractéristiques complexes des ensembles de données que l’on rencontre dans certains domaines d’actualité tel que le Filtrage Collaboratif (FC) et de la fouille de textes. Ces données, souvent représentées sous forme de matrices, sont de très grande dimension (des milliers de variables) et extrêmement creuses (ou sparses, avec plus de 95% de zéros). En plus d’être de grande dimension et sparse, les données rencontrées dans les domaines mentionnés ci-dessus sont également de nature directionnelles. En effet, plusieurs études antérieures ont démontré empiriquement que les mesures directionnelles, telle que la similarité cosinus, sont supérieurs à d’autres mesures, telle que la distance Euclidiennes, pour la classification des documents textuels ou pour mesurer les similitudes entre les utilisateurs/items dans le FC. Cela suggère que, dans un tel contexte, c’est la direction d’un vecteur de données (e.g., représentant un document texte) qui est pertinente, et non pas sa longueur. Il est intéressant de noter que la similarité cosinus est exactement le produit scalaire entre des vecteurs unitaires (de norme 1). Ainsi, d’un point de vue probabiliste l’utilisation de la similarité cosinus revient à supposer que les données sont directionnelles et réparties sur la surface d’une hypersphère unité. En dépit des nombreuses preuves empiriques suggérant que certains ensembles de données sparses et de grande dimension sont mieux modélisés sur une hypersphère unité, la plupart des modèles existants dans le contexte de la fouille de textes et du FC s’appuient sur des hypothèses populaires : distributions Gaussiennes ou Multinomiales, qui sont malheureusement inadéquates pour des données directionnelles. Dans cette thèse, nous nous focalisons sur deux challenges d’actualité, à savoir la classification des documents textuels et la recommandation d’items, qui ne cesse d’attirer l’attention dans les domaines de la fouille de textes et celui du filtrage collaborative, respectivement. Afin de répondre aux limitations ci-dessus, nous proposons une série de nouveaux modèles et algorithmes qui s’appuient sur la distribution de von Mises-Fisher (vMF) qui est plus appropriée aux données directionnelles distribuées sur une hypersphère unité.
Cluster analysis or clustering, which aims to group together similar objects, is undoubtedly a very powerful unsupervised learning technique. With the growing amount of available data, clustering is increasingly gaining in importance in various areas of data science for several reasons such as automatic summarization, dimensionality reduction, visualization, outlier detection, speed up research engines, organization of huge data sets, etc. Existing clustering approaches are, however, severely challenged by the high dimensionality and extreme sparsity of the data sets arising in some current areas of interest, such as Collaborative Filtering (CF) and text mining. Such data often consists of thousands of features and more than 95% of zero entries. In addition to being high dimensional and sparse, the data sets encountered in the aforementioned domains are also directional in nature. In fact, several previous studies have empirically demonstrated that directional measures—that measure the distance between objects relative to the angle between them—, such as the cosine similarity, are substantially superior to other measures such as Euclidean distortions, for clustering text documents or assessing the similarities between users/items in CF. This suggests that in such context only the direction of a data vector (e.g., text document) is relevant, not its magnitude. It is worth noting that the cosine similarity is exactly the scalar product between unit length data vectors, i.e., L 2 normalized vectors. Thus, from a probabilistic perspective using the cosine similarity is equivalent to assuming that the data are directional data distributed on the surface of a unit-hypersphere. Despite the substantial empirical evidence that certain high dimensional sparse data sets, such as those encountered in the above domains, are better modeled as directional data, most existing models in text mining and CF are based on popular assumptions such as Gaussian, Multinomial or Bernoulli which are inadequate for L 2 normalized data. In this thesis, we focus on the two challenging tasks of text document clustering and item recommendation, which are still attracting a lot of attention in the domains of text mining and CF, respectively. In order to address the above limitations, we propose a suite of new models and algorithms which rely on the von Mises-Fisher (vMF) assumption that arises naturally for directional data lying on a unit-hypersphere.
Online marketplaces are able to offer a staggering array of products that no physical store can match. While this makes it more likely for customers to find what they want, in order for online ...providers to ensure a smooth and efficient user experience, they must maintain well-organized catalogs, which depends greatly on the availability of per-product attribute values such as color, material, brand, to name a few. Unfortunately, such information is often incomplete or even missing in practice, and therefore we have to resort to predictive models as well as other sources of information to impute missing attribute values.
In this talk we present the deep learning-based approach that we have developed at Rakuten Group to extract attribute values from product descriptive texts and images. Starting from pretrained architectures to encode textual and visual modalities, we discuss several refinements and improvements that we find necessary to achieve satisfactory performance and meet strict business requirements, namely improving recall while maintaining a high precision (>= 95%). Our methodology is driven by a systematic investigation into several practical research questions surrounding multimodality, which we revisit in this talk. At the heart of our multimodal architecture, is a new method to combine modalities inspired by empirical cross-modality comparisons. We present the latter component in details, point out one of its major limitations, namely exacerbating the issue of modality collapse, i.e., when the model forgets one modality, and describe our mitigation to this problem based on a principled regularization scheme.
We present various empirical results on both Rakuten data as well as public benchmark datasets, which provide evidence of the benefits of our approach compared to several strong baselines. We also share some insights to characterise the circumstances in which the proposed model offers the most significant improvements. We conclude this talk by criticising the current model and discussing possible future developments and improvements.
Our model is successfully deployed in Rakuten Ichiba - a Rakuten marketplace - and we believe that our investigation into multimodal attribute value extraction for e-commerce will benefit other researchers and practitioners alike embarking on similar journeys.
Data clustering and representation learning play an indispensable role in data science. They are very useful to explore massive data in many fields, including information retrieval, natural language ...processing, bioinformatics, recommender systems, and computer vision. Despite their success, most existing clustering methods are severely challenged by the data generated by modern applications, which are typically high dimensional, noisy, heterogeneous, and sparse or even collected from multiple sources or represented by multiple views where each describes a perspective of the data. This has driven many researchers to investigate new effective clustering models to overcome these difficulties. One promising category of such models relies on representation learning. Indeed, learning a good data representation is crucial for clustering algorithms, and combining the two tasks is a common way of exploring this type of data. The idea is to embed the original data into a low dimensional latent space and then perform clustering on this new space. However, both tasks can be carried out sequentially or jointly. Many clustering algorithms, including deep learning versions, are based on these two modes of combining the two tasks.