The typical offline protocol to evaluate recommendation algorithms is to collect a dataset of user-item interactions and then use a part of this dataset to train a model, and the remaining data to measure how closely the model recommendations match the observed user interactions. This protocol is straightforward, useful and practical, but it only provides snapshot performance. We know, however, that online systems evolve over time. In general, it is a good idea that models are frequently retrained with recent data. But if this is the case, to what extent can we trust previous evaluations? How will a model perform when a different pattern (re)emerges? In this paper we propose a methodology to study how recommendation models behave when they are retrained. The idea is to profile algorithms according to their ability to, on the one hand, retain past patterns – stability – and, on the other hand, (quickly) adapt to changes – plasticity. We devise an offline evaluation protocol that provides detail on the long-term behavior of models, and that is agnostic to datasets, algorithms and metrics. To illustrate the potential of this framework, we present preliminary results of three different types of algorithms on the GoodReads dataset that suggest different stability and plasticity profiles depending on the algorithmic technique, and a possible trade-off between stability and plasticity. We further discuss the potential and limitations of the proposal and advance some possible improvements.

We introduce Fed-VFDT, a federated adaptation of the Very Fast Decision Tree (VFDT) algorithm for classification over streaming data. While VFDT is a widely adopted online learning algorithm, its sequential and order-sensitive nature poses challenges in federated settings, marked by statistical heterogeneity and communication constraints. Fed-VFDT addresses these issues by having each client incrementally train a local VFDT and report split statistics to a central server when a leaf satisfies the Hoeffding criterion. The server selects a global splitting feature by aggregating clients' proposals according to a configurable strategy: quorum, merit-based selection, or majority voting. Once a feature is selected, it is broadcast to all clients, which apply the split at the corresponding tree path using their locally computed thresholds. We evaluate Fed-VFDT against its centralized counterpart using predictive and structural metrics, demonstrating that it maintains comparable performance while reducing communication and preserving synchronized tree growth.

The European Union (EU) Digital Services Act (DSA) has introduced a novel set of rules for online platforms and search engines, with significant implications for the Recommender Systems community. Through its data access mechanisms, the DSA invites researchers to request both publicly available and private data from Very Large Online Platforms (VLOPs) and Very Large Search Engines (VLOSEs) – those with more than 45 million active recipients in the EU – to investigate systemic risks associated with the dissemination of illegal content, risks to the exercise of fundamental rights, and negative effects on electoral processes, public health, and gender-based violence. This tutorial is aimed at researchers who are interested in submitting such data access requests and will provide them with the knowledge to do so by introducing the relevant definitions and provisions of the DSA, and addressing the most important procedural steps to obtain data access and will provide attendees with a comprehensive understanding of the DSA’s data access implications for RecSys research. The tutorial targets researchers, practitioners, and students in understanding current developments in online platform regulation in Europe and their impact on RecSys research.

Effective anomaly detection in telecommunication networks is essential for securing digital transactions and supporting the sustainability of our global information ecosystem. However, the volume of data in such high-speed distributed environments imposes strict latency and scalability requirements on anomaly detection systems. This study focuses on distributed heavy hitter detection in telephone networks – a critical component of network traffic analysis and fraud detection. We propose a federated version of the Lossy Counting algorithm and compare it to its centralized version. Our experimental results reveal that the federated approach can detect considerably more unique heavy hitters than the centralized method while enhancing privacy. Furthermore, Federated Lossy Counting does not need a large amount of centralized processing power since it can leverage the networked infrastructure with minimal impact on bandwidth and computing power.

Modern online platforms for user modeling and recommendation require complex data infrastructures to collect and process data. Some of this data has to be kept to later be used in batches to train personalization models. However, since user activity data can be generated at very fast rates it is also useful to have algorithms able to process data streams online, in real time. Given the continuous and potentially fast change of content, context and user preferences or intents, stream-based models, and their synchronization with batch models can be extremely challenging. Therefore, it is important to investigate methods able to transparently and continuously adapt to the inherent dynamics of user interactions, preferably over long periods of time. Models able to continuously learn from such flows of data are gaining attention in the recommender systems community, and are being increasingly deployed in online platforms. However, many challenges associated with learning from streams need further investigation.

Tor is the most popular anonymity network in the world. It relies on advanced security and obfuscation techniques to ensure the privacy of its users and free access to the Internet. However, the investigation of traffic correlation attacks against Tor Onion Services (OSes) has been relatively overlooked in the literature. In particular, determining whether it is possible to emulate a global passive adversary capable of deanonymizing the IP addresses of both the Tor OSes and of the clients accessing them has remained, so far, an open question. In this paper, we present ongoing work toward addressing this question and reveal some preliminary results on a scalable traffic correlation attack that can potentially be used to deanonymize Tor OS sessions. Our attack is based on a distributed architecture involving a group of colluding ISPs from across the world. After collecting Tor traffic samples at multiple vantage points, ISPs can run them through a pipeline where several stages of traffic classifiers employ complementary techniques that result in the deanonymization of OS sessions with high confidence (i.e., low false positives). We have responsibly disclosed our early results with the Tor Project team and are currently working not only on improving the effectiveness of our attack but also on developing countermeasures to preserve Tor users' privacy.

Modern online systems for user modeling and recommendation need to continuously deal with complex data streams generated by users at very fast rates. This can be overwhelming for systems and algorithms designed to train recommendation models in batches, given the continuous and potentially fast change of content, context and user preferences or intents. Therefore, it is important to investigate methods able to transparently and continuously adapt to the inherent dynamics of user interactions, preferably for long periods of time. Online models that continuously learn from such flows of data are gaining attention in the recommender systems community, given their natural ability to deal with data generated in dynamic, complex environments. User modeling and personalization can particularly benefit from algorithms capable of maintaining models incrementally and online.

Modern online services continuously generate data at very fast rates. This continuous flow of data encompasses content – e.g. posts, news, products, comments –, but also user feedback – e.g. ratings, views, reads, clicks –, together with context data – user device, spacial or temporal data, user task or activity, weather. This can be overwhelming for systems and algorithms designed to train in batches, given the continuous and potentially fast change of content, context and user preferences or intents. Therefore, it is important to investigate online methods able to transparently adapt to the inherent dynamics of online services. Incremental models that learn from data streams are gaining attention in the recommender systems community, given their natural ability to deal with the continuous flows of data generated in dynamic, complex environments. User modeling and personalization can particularly benefit from algorithms capable of maintaining models incrementally and online.

Although widely used, the majority of current music recommender systems still focus on recommendations' accuracy, user preferences and isolated item characteristics, without evaluating other important factors, like the joint item selections and the recommendation moment. However, when it comes to playlist recommendations, additional dimensions, as well as the notion of user experience and perception, should be taken into account to improve recommendations' quality. In this work, HybA, a hybrid recommender system for automatic playlist continuation, that combines Latent Dirichlet Allocation and Case-Based Reasoning, is proposed. This system aims to address “similar concepts” rather than similar users. More than generating a playlist based on user requirements, like automatic playlist generation methods, HybA identifies the semantic characteristics of a started playlist and reuses the most similar past ones, to recommend relevant playlist continuations. In addition, support to beyond accuracy dimensions, like increased coherence or diverse items' discovery, is provided. To overcome the semantic gap between music descriptions and user preferences, identify playlist structures and capture songs' similarity, a graph model is used. Experiments on real datasets have shown that the proposed algorithm is able to outperform other state of the art techniques, in terms of accuracy, while balancing between diversity and coherence.

Online incremental models for recommendation are nowadays pervasive in both the industry and the academia. However, there is not yet a standard evaluation methodology for the algorithms that maintain such models. Moreover, online evaluation methodologies available in the literature generally fall short on the statistical validation of results, since this validation is not trivially applicable to stream-based algorithms. We propose ak-fold validation framework for the pairwise comparison of recommendation algorithms that learn from user feedback streams, using prequential evaluation. Our proposal enables continuous statistical testing on adaptive-size sliding windows over the outcome of the prequential process, allowing practitioners and researchers to make decisions in real time based on solid statistical evidence. We present a set of experiments to gain insights on the sensitivity and robustness of two statistical tests-McNemar's and Wilcoxon signed rank-in a streaming data environment. Our results show that besides allowing a real-time, fine-grained online assessment, the online versions of the statistical tests are at least as robust as the batch versions, and definitely more robust than a simple prequential single-fold approach.

Modern online web-based systems continuously generate data at very fast rates. This continuous flow of data encompasses web content – e.g. posts, news, products, comments –, but also user feedback – e.g. ratings, views, reads, clicks, thumbs up –, as well as context information – device used, geographic info, social network, current user activity, weather. This is potentially overwhelming for systems and algorithms design to train in offline batches, given the continuous and potentially fast change of content, context and user preferences. Therefore it is important to investigate online methods to be able to transparently adapt to the inherent dynamics of online systems. Incremental models that learn from data streams are gaining attention in the recommender systems community, given their natural ability to deal with data generated in dynamic, complex environments. User modeling and personalization can particularly benefit from algorithms capable of maintaining models incrementally and online, as data is generated.

The ever-growing nature of user generated data in online systems poses obvious challenges on how we process such data. Typically, this issue is regarded as a scalability problem and has been mainly addressed with distributed algorithms able to train on massive amounts of data in short time windows. However, data is inevitably adding up at high speeds. Eventually one needs to discard or archive some of it. Moreover, the dynamic nature of data in user modeling and recommender systems, such as change of user preferences, and the continuous introduction of new users and items make it increasingly difficult to maintain up-to-date, accurate recommendation models. The objective of this workshop is to bring together researchers and practitioners interested in incremental and adaptive approaches to stream-based user modeling, recommendation and personalization, including algorithms, evaluation issues, incremental content and context mining, privacy and transparency, temporal recommendation or software frameworks for continuous learning.

Abstract Ensemble methods have been successfully used in the past to improve recommender systems; however, they have never been studied with incremental recommendation algorithms. Many online recommender systems deal with continuous, potentially fast, and unbounded flows of data—big data streams—and often need to be responsive to fresh user feedback, adjusting recommendations accordingly. This is clear in tasks such as social network feeds, news recommender systems, automatic playlist completion, and other similar applications. Batch ensemble approaches are not suitable to perform continuous learning, given the complexity of retraining new models on demand. In this paper, we adapt a general purpose online bagging algorithm for top‐N recommendation tasks and propose two novel online bagging methods specifically tailored for recommender systems. We evaluate the three approaches, using an incremental matrix factorization algorithm for top‐N recommendation with positive‐only user feedback data as the base model. Our results show that online bagging is able to improve accuracy up to 55% over the baseline, with manageable computational overhead.