Efficient Learning of Sparse Representations from Interactions
Authors
Vančura, V.; Spišák, M.; da Silva Alves, R.; Peška, L.
Year
2026
Published
Proceedings of the ACM Web Conference 2026. New York: ACM, 2026. p. 8577-8580. ISBN 979-8-4007-2307-0.
Type
Proceedings paper
Departments
Annotation
Behavioral patterns captured in embeddings learned from interaction data are pivotal across various stages of production recommender systems. However, in the initial retrieval stage, practitioners face an inherent tradeoff between embedding expressiveness and the scalability and latency of serving components, resulting in the need for representations that are both compact and expressive. To address this challenge, we propose a training strategy for learning high-dimensional sparse embedding layers in place of conventional dense ones, balancing efficiency, representational expressiveness, and interpretability. To demonstrate our approach, we modified the production-grade collaborative filtering autoencoder ELSA, achieving up to 10× reduction in embedding size with no loss of recommendation accuracy, and up to 100× reduction with only a 2.5% loss. Moreover, the active embedding dimensions reveal an interpretable inverted-index structure that segments items in a way directly aligned with the model's latent space, thereby enabling integration of segment-level recommendation functionality (e.g., 2D homepage layouts) within the candidate retrieval model itself. Source codes, additional results, as well as a live demo are available at https://github.com/zombak79/compressed\_elsa
Learning Minimally Rigid Graphs with High Realization Counts
Authors
Year
2026
Published
Proceedings of the 35th International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization, 2026. p. 4813-4821. ISBN 978-1-956792-09-6.
Type
Proceedings paper
Departments
Annotation
For minimally rigid graphs, the same edge-length data can admit multiple realizations (up to translations and rotations). Finding graphs with exceptionally many realizations is an extremal problem in rigidity theory, but exhaustive search quickly becomes infeasible due to the super-exponential growth of the number of candidate graphs and the high cost of realization-count evaluation. We propose a reinforcement-learning approach that constructs minimally rigid graphs via 0- and 1-extensions, also known as Henneberg moves. We optimize realization-count invariants using the Deep Cross-Entropy Method with a policy parameterized by a Graph Isomorphism Network encoder and a permutation-equivariant extension-level action head. Empirically, our method matches the known optima for planar realization counts and improves the best known bounds for spherical realization counts, yielding new record graphs.
Reasoning-Grounded Natural Language Explanations for Language Models
Authors
Year
2026
Published
Explainable Artificial Intelligence, Third World Conference, xAI 2025, Istanbul, Turkey, July 9–11, 2025, Proceedings, Part III. Cham: Springer, 2026. p. 3-18. Communications in Computer and Information Science. vol. 2578. ISSN 1865-0929. ISBN 978-3-032-08326-5.
Type
Proceedings paper
Departments
Annotation
We propose a large language model explainability technique for obtaining faithful natural language explanations by grounding the explanations in a reasoning process. When converted to a sequence of tokens, the outputs of the reasoning process can become part of the model context and later be decoded to natural language as the model produces either the final answer or the explanation. To improve the faithfulness of the explanations, we propose to use a joint predict-explain approach, in which the answers and explanations are inferred directly from the reasoning sequence, without the explanations being dependent on the answers and vice versa. We demonstrate the plausibility of the proposed technique by achieving a high alignment between answers and explanations in several problem domains, observing that language models often simply copy the partial decisions from the reasoning sequence into the final answers or explanations. Furthermore, we show that the proposed use of reasoning can also improve the quality of the answers.
Active Recommendation for Email Outreach Dynamics
Authors
Year
2025
Published
ACM International Conference on Information and Knowledge Management. New York: ACM, 2025. p. 5540-5544. 34. ISBN 979-8-4007-2040-6.
Type
Proceedings paper
Departments
Annotation
Email outreach remains a cornerstone of modern marketing, enabling direct, timely communication. However, this strategy faces significant personalization challenges, since new campaigns typically lack historical interaction data and rich side information. In this work, we propose a framework that combines collaborative-filtering (CF) signals derived from a shallow autoencoder (SAE) with a Thompson Sampling-based multi-armed bandit to dynamically select small batches of recipients for each email template. We show SAEs help balance exploration and exploitation by quantifying recipient informativeness and confidence, enabling efficient personalization without retraining during active learning. To facilitate reproducibility and future research, we release a large dataset of almost 15 million recipient-message interactions, offering new insights into email outreach dynamics for CF. Our experiments show that our method outperforms multiple baselines in retrieval metrics while retaining interpretable model components.
Context-Aware REpresentation: Jointly Learning Item Features and Selection From Triplets
Authors
da Silva Alves, R.; Ledent, A.
Year
2025
Published
IEEE Transactions on Neural Networks and Learning Systems. 2025, 36(4), 6492-6502. ISSN 2162-237X.
Type
Article
Departments
Annotation
In areas of machine learning such as cognitive modeling or recommendation, user feedback is usually context-dependent. For instance, a website might provide a user with a set of recommendations and observe which (if any) of the links were clicked by the user. Similarly, there is growing interest in the so-called “odd-one-out” learning setting, where human participants are provided with a basket of items and asked which is the most dissimilar to the others. In both of those cases, the presence of all the items in the basket can influence the final decision. In this article, we consider a classification task where each input consists of three items (a triplet), and the task is to predict which of the three will be selected. Our aim is not only to return accurate predictions for the selection task, but also to additionally provide interpretable feature representations for both the context and for each individual item. To achieve this, we introduce CARE, a specialized neural network architecture that yields Context-Aware REpresentations of items based on observations of triplets of items alone. We demonstrate that, in addition to achieving state-of-the-art performance at the selection task, our model can produce meaningful representations both for each item, as well for each context (triplet of items). This is done using only triplet responses: CARE has no access to supervised item-level information. In addition, we prove parameter counting generalization bounds for our model in the i.i.d. setting, demonstrating that the apparent sample sparsity arising from the combinatorially large number of possible triplets is no obstacle to efficient learning.
Conv4Rec: A 1-by-1 Convolutional Autoencoder for User Profiling Through Joint Analysis of Implicit and Explicit Feedback
Authors
Ledent, A.; Kasalický, P.; da Silva Alves, R.; Lauw, H.W.
Year
2025
Published
IEEE Transactions on Neural Networks and Learning Systems. 2025, 36(12), 20035-20049. ISSN 2162-237X.
Type
Article
Departments
Annotation
We introduce a new convolutional autoencoder architecture for user modeling and recommendation tasks with several improvements over the state of the art. First, our model has the flexibility to learn a set of associations and combinations between different interaction types in a way that carries over to each user and item. Second, our model is able to learn jointly from both the explicit ratings and the implicit information in the sampling pattern (which we refer to as ”implicit feedback”). It can also make separate predictions for the probability of consuming content and the likelihood of granting it a high rating if observed. This not only allows the model to make predictions for both the implicit and explicit feedback, but also increases the informativeness of the predictions: in particular, our model can identify items that users would not have been likely to consume naturally, but would be likely to enjoy if exposed to them. Finally, we provide several generalization bounds for our model, which, to the best of our knowledge, are among the first generalization bounds for autoencoders in a Recommender systems setting; we also show that optimizing our loss function guarantees the recovery of the exact sampling distribution over interactions up to a small error in total variation. In experiments on several real-life datasets, we achieve state-of-the-art performance on both the implicit and explicit feedback prediction tasks despite relying on a single model for both, and benefiting from additional interpretability in the form of individual predictions for the probabilities of each possible rating.
From Representation to Response: Assessing the Alignment of Large Language Models with Human Judgment Patterns
Authors
Hrytsyna, A.; da Silva Alves, R.
Year
2025
Published
ACM Transactions on Intelligent Systems and Technology. 2025, 16(6), 1-23. ISSN 2157-6904.
Type
Article
DOI
Annotation
Large language models (LLMs) are sophisticated artificial intelligence systems designed to process and understand natural language at a complex level. The recent progress of these models, culminating in chat-based LLMs, has democratized the accessibility of these sophisticated intelligent systems, showcasing how machine learning methods can help humans in daily tasks. This research addresses the growing interest in understanding the mechanisms of LLMs and in evaluating their alignment with human cognition. We introduce an innovative alignment assessment strategy in the realm of LLMs that diverges from traditional approaches, utilizing the odd-one-out triplet-based task to investigate the alignment of LLMs' representations with human object concept mental organization. Our methodology, which incorporates image captioning and zero/few-shot learning accuracy scoring, is designed to evaluate language models' ability to predict similarities and differences in object concepts. A comprehensive experimental evaluation was conducted, involving four captioning strategies, twenty-four LLMs across eight model families, and three scoring procedures, utilizing a significantly large dataset for enhanced understanding of LLM comprehensibility. Finally, our study explores the impact of object description comprehensiveness on model-human representation alignment and analyzes a subset of randomly selected triplets to assess how LLMs are able to represent different levels of human judgment patterns.
Generalization Bounds for Rank-sparse Neural Networks
Authors
Ledent, A.; da Silva Alves, R.; Lei, Y.
Year
2025
Published
Advances in Neural Information Processing Systems 38 (NeurIPS 2025). Neural Information Processing Systems Foundation, Inc., 2025. p. 1-70. vol. 38. ISSN 1049-5258.
Type
Proceedings paper
Departments
Annotation
It has been recently observed in much of the literature that neural networks exhibit a bottleneck rank property: for larger depths, the activation and weights of neural networks trained with gradient-based methods tend to be of approximately low rank. In fact, the rank of the activations of each layer converges to a fixed value referred to as the "bottleneck rank", which is the minimum rank required to represent the training data. This perspective is in line with the observation that regularizing linear networks (without activations) with weight decay is equivalent to minimizing the Schatten
quasi norm of the neural network. In this paper we investigate the implications of this phenomenon for generalization. More specifically, we prove generalization bounds for neural networks which exploit the approximate low rank structure of the weight matrices if present. The final results rely on the Schatten quasi norms of the weight matrics. As
increases, the bound behaves more like a norm-based bound instead. The proof techniques involve a careful interpolation between the parametric and norm based regimes. We also demonstrate in experiments that this bound outperforms both classic parameter counting and norm based bounds in the typical overparametrized regime.
Multitask learning for cognitive sciences triplet analysis
Authors
Stambrouski, T.; da Silva Alves, R.
Year
2025
Published
Expert Systems with Applications. 2025, 267 1-10. ISSN 0957-4174.
Type
Article
Annotation
The triplet-based odd-one-out problem, which involves trials where human subjects are asked to select the most different concept among three, is a well-studied task in cognitive sciences. With the release of a large triplet-based dataset, THINGS, there has been a recent surge in the popularity of machine learning models aimed at learning mathematical representations of object concepts, such as SPoSE, VICE, and CARE. The first two models learn representations by maximizing the similarity between the two most similar objects, while the latter diverges by directly learning the odd-one-out, making its embedding more distant. No prior attempts have integrated both paradigms, which are important for understanding object representation in cognitive science. In this paper, we propose MASTER, a multitask learning method for the triplet problem that encapsulates both paradigms. Our results demonstrate that our method not only better predicts the odd-one-out object but also provides insightful representations for studying these concepts. Furthermore, we studied the conditions under which each model performs better, offering valuable insights for future research on how these paradigms affect human understanding of object concepts.
Probabilistic Modeling, Learnability and Uncertainty Estimation for Interaction Prediction in Movie Rating Datasets
Authors
Poernomo, J.; Tan, N.G.L.; da Silva Alves, R.; Ledent, A.
Year
2025
Published
RecSys '25: Proceedings of the Nineteenth ACM Conference on Recommender Systems. New York: ACM, 2025. p. 1261-1266. ISBN 979-8-4007-1364-4.
Type
Proceedings paper
Departments
Annotation
In this paper, we examine the hypothesis that the interactions recorded in many Recommendation Systems datasets are distributed according to a low-rank distribution, i.e. a mixture of factorizable distributions. Surprisingly, we find that on several popular datasets, a simple non-negative matrix factorization method equals or outperforms more modern methods such as LightGCN, which indicates that the sampling distribution over interactions is indeed low-rank. Furthermore, we mathematically prove that low-rank distributions are learnable with a sparse number of observations (where m/n and r refer to the number of users/items and the non-negative rank respectively) both in terms of the total variation norm and in terms of the expected recall at k, arguably providing some of the first generalization bounds for recommender systems in the implicit feedback setting. We also provide a modified version of the NMF algorithm which provides further performance improvements compared to the standard NMF baseline on the smaller datasets considered. Finally, we propose the theoretically grounded concept of empirical expected recall as an uncertainty estimate for probabilistic models of the recommendation task, and demonstrate its success in a setting where user-wise abstentions are allowed.
Recurrent Autoregressive Linear Model for Next-Basket Recommendation
Authors
Zmeškalová, T.; Ledent, A.; Spisak, M.; Kordík, P.; da Silva Alves, R.
Year
2025
Published
RecSys '25: Proceedings of the Nineteenth ACM Conference on Recommender Systems. New York: ACM, 2025. p. 1273-1278. ISBN 979-8-4007-1364-4.
Type
Proceedings paper
Departments
Annotation
Next-basket recommendation aims to predict the (sets of) items that a user is most likely to purchase during their next visit, capturing both short-term sequential patterns and long-term user preferences. However, effectively modeling these dynamics remains a challenge for traditional methods, which often struggle with interpretability and computational efficiency, particularly when dealing with intricate temporal dependencies and inter-item relationships. In this paper, we propose ReALM, a Recurrent Autoregressive Linear Model that explicitly captures temporal item-to-item dependencies across multiple time steps. By leveraging a recurrent loss function and a closed-form optimization solution, our approach offers both interpretability and scalability while maintaining competitive accuracy. Experimental results on real-world datasets demonstrate that ReALM outperforms several state-of-the-art baselines in both recommendation quality and efficiency, offering a robust and interpretable solution for modern personalization systems.
SCORE: A convolutional approach for football event forecasting
Authors
Year
2025
Published
International Journal of Forecasting. 2025, 41(4), 1636-1652. ISSN 0169-2070.
Type
Article
Departments
Annotation
Football (also known as soccer or association football) is the most popular sport in the world. It is a blend of skill and luck, making it highly unpredictable. To address this unpredictability, there has been a surge in popularity over the past decade in employing machine learning techniques for forecasting football-related features. This trend aligns with the growing professionalism in football analytics. Despite this progress, the existing body of work remains in its early stages, lacking the depth required to capture the intricate nuances of the sport. In this study, we introduce a convolutional approach designed to predict the occurrence of the next event in a football match, such as a goal or a corner kick, relying solely on easy-to-access past events for predictions. Our methodology adopts an online approach, meaning predictions can be computed during a live match. To validate our approach, we conduct a comprehensive evaluation against five baseline models, utilizing data from various elite European football leagues. Additionally, an ablation study is performed to understand the underlying mechanisms of our method. Finally, we present practical applications and interpretable aspects of our proposed approach.
The Future is Sparse: Embedding Compression for Scalable Retrieval in Recommender Systems
Authors
Year
2025
Published
RecSys '25: Proceedings of the Nineteenth ACM Conference on Recommender Systems. New York: ACM, 2025. p. 1099-1103. ISBN 979-8-4007-1364-4.
Type
Proceedings paper
Departments
Annotation
Industry-scale recommender systems face a core challenge: representing entities with high cardinality, such as users or items, using dense embeddings that must be accessible during both training and inference. However, as embedding sizes grow, memory constraints make storage and access increasingly difficult. We describe a lightweight, learnable embedding compression technique that projects dense embeddings into a high-dimensional, sparsely activated space. Designed for retrieval tasks, our method reduces memory requirements while preserving retrieval performance, enabling scalable deployment under strict resource constraints. Our results demonstrate that leveraging sparsity is a promising approach for improving the efficiency of large-scale recommenders. We release our code at https://github.com/recombee/CompresSAE.
Generalization Analysis of Deep Non-linear Matrix Completion
Authors
Ledent, A.; da Silva Alves, R.
Year
2024
Published
Proceedings of Machine Learning Research. Proceedings of Machine Learning Research, 2024. p. 26290-26360. vol. 235. ISSN 2640-3498.
Type
Proceedings paper
Departments
Annotation
We provide generalization bounds for matrix completion with Schatten 𝑝
quasi-norm constraints, which is equivalent to deep matrix factorization with Frobenius constraints. In the uniform sampling regime, the sample complexity scales like 𝑂˜(𝑟𝑛)
where 𝑛 is the size of the matrix and 𝑟 is a constraint of the same order as the ground truth rank in the isotropic case. In the distribution-free setting, the bounds scale as 𝑂˜(𝑟1−𝑝2𝑛1+𝑝2), which reduces to the familiar 𝑟√𝑛32 for 𝑝=1 . Furthermore, we provide an analogue of the weighted trace norm for this setting which brings the sample complexity down to 𝑂˜(𝑛𝑟) in all cases. We then present a non-linear model, Functionally Rescaled Matrix Completion (FRMC) which applies a single trainable function from ℝ→ℝ
to each entry of a latent matrix, and prove that this adds only negligible terms of the overall sample complexity, whilst experiments demonstrate that this simple model improvement already leads to significant gains on real data. We also provide extensions of our results to various neural architectures, thereby providing the first comprehensive uniform convergence PAC analysis of neural network matrix completion.
Recommendations with minimum exposure guarantees: A post-processing framework
Authors
Lopes, R.; da Silva Alves, R.; Ledent, A.; Santos, R.L.T.; Kloft, M.
Year
2024
Published
Expert Systems with Applications. 2024, 236 1-9. ISSN 1873-6793.
Type
Article
Departments
Annotation
Relevance-based ranking is a popular ingredient in recommenders, but it frequently struggles to meet fairness criteria because social and cultural norms may favor some item groups over others. For instance, some items might receive lower ratings due to some sort of bias (e.g. gender bias). A fair ranking should balance the exposure of items from advantaged and disadvantaged groups. To this end, we propose a novel post-processing framework to produce fair, exposure-aware recommendations. Our approach is based on an integer linear programming model maximizing the expected utility while satisfying a minimum exposure constraint. The model has fewer variables than previous work and thus can be deployed to larger datasets and allows the organization to define a minimum level of exposure for groups of items. We conduct an extensive empirical evaluation indicating that our new framework can increase the exposure of items from disadvantaged groups at a small cost of recommendation accuracy
Regionalization-Based Collaborative Filtering: Harnessing Geographical Information in Recommenders
Authors
Year
2024
Published
ACM Transactions on Spatial Algorithms and Systems. 2024, 10(2), 1-23. ISSN 2374-0361.
Type
Article
DOI
Departments
Annotation
Regionalization, also known as spatially constrained clustering, is an unsupervised machine learning technique used to identify and define spatially contiguous regions. In this work, we introduce a methodology to regionalize recommendation systems (RSs) based on a collaborative filtering approach. Two main challenges arise when performing regionalization based on users' preferences in RSs: (1) unstructured data, as interactions are often scarce and observed on a smaller scale; and (2) the difficulty of evaluation of the quality of the clustering results. To address these challenges, our methodology relies on inductive matrix completion (IMC), a fundamental approach to recover unknown entries of a rating matrix while utilizing region information to extract a region-based feature matrix. With this feature matrix, our method becomes adaptive and seamlessly integrates with various regionalization algorithms to create regionalization candidates. This enables us to derive more accurate recommendations that consider regionalized effects and discover interesting patterns in localized user behavior. We experimentally evaluate our model on synthetic datasets to demonstrate its efficacy in settings where our underlying assumptions are correct. Furthermore, we present a real-world case study illustrating the interpretable information the model can derive in terms of regionalized recommendation relevance.
Uncertainty-Adjusted Recommendation via Matrix Factorization With Weighted Losses
Authors
da Silva Alves, R.; Ledent, A.; Kloft, M.
Year
2024
Published
IEEE Transactions on Neural Networks and Learning Systems. 2024, 35(11), 15624-15637. ISSN 2162-237X.
Type
Article
Departments
Annotation
In a recommender systems (RSs) dataset, observed ratings are subject to unequal amounts of noise. Some users might be consistently more conscientious in choosing the ratings they provide for the content they consume. Some items may be very divisive and elicit highly noisy reviews. In this article, we perform a nuclear-norm-based matrix factorization method which relies on side information in the form of an estimate of the uncertainty of each rating. A rating with a higher uncertainty is considered more likely to be erroneous or subject to large amounts of noise, and therefore more likely to mislead the model. Our uncertainty estimate is used as a weighting factor in the loss we optimize. To maintain the favorable scaling and theoretical guarantees coming with nuclear norm regularization even in this weighted context, we introduce an adjusted version of the trace norm regularizer which takes the weights into account. This regularization strategy is inspired from the weighted trace norm which was introduced to tackle nonuniform sampling regimes in matrix completion. Our method exhibits state-of-the-art performance on both synthetic and real life datasets in terms of various performance measures, confirming that we have successfully used the auxiliary information extracted.
Unraveling the Dynamics of Stable and Curious Audiences in Web Systems
Authors
da Silva Alves, R.; Ledent, A.; Assunção, R.; Vaz-De-Melo, P.; Kloft, M.
Year
2024
Published
Proceedings of the ACM Web Conference 2024. New York: ACM, 2024. p. 2464-2475. ISBN 979-8-4007-0171-9.
Type
Proceedings paper
Departments
Annotation
We propose the Burst-Induced Poisson Process (BPoP), a model designed to analyze time series data such as feeds or search queries. BPoP can distinguish between the slowly-varying regular activity of a stable audience and the bursty activity of a curious audience, often seen in viral threads. Our model consists of two hidden, interacting processes: a self-feeding process (SFP) that generates bursty behavior related to viral threads, and a non-homogeneous Poisson process (NHPP) with step function intensity that is influenced by the bursts from the SFP. The NHPP models the normal background behavior, driven solely by the overall popularity of the topic among the stable audience. Through extensive empirical work, we have demonstrated that our model fits and characterizes a large number of real datasets more effectively than state-of-the-art models. Most importantly, BPoP can quantify the stable audience of media channels over time, serving as a valuable indicator of their popularity.
Bridging Offline-Online Evaluation with a Time-dependent and Popularity Bias-free Offline Metric for Recommenders
Authors
Year
2023
Published
Proceedings of EvalRS: A Rounded Evaluation Of Recommender Systems 2023. CEUR-WS.org, 2023. ISSN 1613-0073.
Type
Proceedings paper
Departments
Annotation
The evaluation of recommendation systems is a complex task. The offline and online evaluation metrics for recommender systems are ambiguous in their true objectives. The majority of recently published papers benchmark their methods using ill-posed offline evaluation methodology that often fails to predict true online performance. Because of this, the impact that academic research has on the industry is reduced. The aim of our research is to investigate and compare the online performance of offline evaluation metrics. We show that penalizing popular items and considering the time of transactions during the evaluation significantly improves our ability to choose the best recommendation model for a live recommender system. Our results, averaged over five large-size real-world live data procured from recommenders, aim to help the academic community to understand better offline evaluation and optimization criteria that are more relevant for real applications of recommender systems.
Generalization Bounds for Inductive Matrix Completion in Low-noise Settings
Authors
Ledent, A.; da Silva Alves, R.; Lei, Y.; Guermeur, Y.; Kloft, M.
Year
2023
Published
Proceedings of the 37th AAAI Conference on Artificial Intelligence. Menlo Park: AAAI Press, 2023. p. 8447-8455. vol. 37. ISSN 2374-3468. ISBN 978-1-57735-880-0.
Type
Proceedings paper
Departments
Annotation
We study inductive matrix completion (matrix completion with side information) under an i.i.d. subgaussian noise assumption at a low noise regime, with uniform sampling of the entries. We obtain for the first time generalization bounds with the following three properties: (1) they scale like the standard deviation of the noise and in particular approach zero in the exact recovery case; (2) even in the presence of noise, they converge to zero when the sample size approaches infinity; and (3) for a fixed dimension of the side information, they only have a logarithmic dependence on the size of the matrix. Differently from many works in approximate recovery, we present results both for bounded Lipschitz losses and for the absolute loss, with the latter relying on Talagrand-type inequalities. The proofs create a bridge between two approaches to the theoretical analysis of matrix completion, since they consist in a combination of techniques from both the exact recovery literature and the approximate recovery literature.
Uncertainty-adjusted Inductive Matrix Completion with Graph Neural Networks
Authors
Kasalický, P.; Ledent, A.; da Silva Alves, R.
Year
2023
Published
RecSys '23: Proceedings of the 17th ACM Conference on Recommender Systems. New York: Association for Computing Machinery, 2023. p. 1169-1174. ISBN 979-8-4007-0241-9.
Type
Proceedings paper
Departments
Annotation
We propose a robust recommender systems model which performs matrix completion and a ratings-wise uncertainty estimation jointly. Whilst the prediction module is purely based on an implicit low-rank assumption imposed via nuclear norm regularization, our loss function is augmented by an uncertainty estimation module which learns an anomaly score for each individual rating via a Graph Neural Network: data points deemed more anomalous by the GNN are downregulated in the loss function used to train the low-rank module. The whole model is trained in an end-to-end fashion, allowing the anomaly detection module to tap on the supervised information available in the form of ratings. Thus, our model’s predictors enjoy the favourable generalization properties that come with being chosen from small function space (i.e., low-rank matrices), whilst exhibiting the robustness to outliers and flexibility that comes with deep learning methods. Furthermore, the anomaly scores themselves contain valuable qualitative information. Experiments on various real-life datasets demonstrate that our model outperforms standard matrix completion and other baselines, confirming the usefulness of the anomaly detection module.
Scalable Linear Shallow Autoencoder for Collaborative Filtering
Authors
Year
2022
Published
RecSys '22: Proceedings of the 16th ACM Conference on Recommender Systems. New York: Association for Computing Machinery, 2022. p. 604-609. ISBN 978-1-4503-9278-5.
Type
Proceedings paper
Departments
Annotation
Recently, the RS research community has witnessed a surge in popularity for shallow autoencoder-based CF methods. Due to its straightforward implementation and high accuracy on item retrieval metrics, EASE is potentially the most prominent of these models. Despite its accuracy and simplicity, EASE cannot be employed in some real-world recommender system applications due to its inability to scale to huge interaction matrices. In this paper, we proposed ELSA, a scalable shallow autoencoder method for implicit feedback recommenders. ELSA is a scalable autoencoder in which the hidden layer is factorizable into a low-rank plus sparse structure, thereby drastically lowering memory consumption and computation time. We conducted a comprehensive offline experimental section that combined synthetic and several real-world datasets. We also validated our strategy in an online setting by comparing ELSA to baselines in a live recommender system using an A/B test. Experiments demonstrate that ELSA is scalable and has competitive performance. Finally, we demonstrate the explainability of ELSA by illustrating the recovered latent space.