Aris Anagnostopoulos

The application of offline reinforcement learning (RL) in recommendation systems has proven valuable in improving user engagement. However, a significant challenge in real-world scenarios is the delayed nature of user feedback, which can degrade algorithm performance and complicate the task of reducing user churn, i.e., losing clients to competitors. In this article, we present an end-to-end framework called Churn Reduction Agent-Based (CRAB), which is specifically designed to efficiently incorporate delayed feedback into the recommendation process. CRAB uses a conservative Q-learning model to recommend offers to customers, incorporating delayed churn feedback directly into the reward function of the RL model. This allows the RL agent to optimize recommendations not only based on immediate acceptance probabilities, but also considering long-term user retention. Our approach also uses additional auxiliary model predictions to improve the representation of customer states in the RL environment, which significantly improves the framework’s performance. We validate our method using large-scale proprietary data from a real-world telecommunication company, addressing the complexities of scalability and practical implementation. Extensive offline experiments show that CRAB outperforms existing methods, including both internal company policies and state-of-the-art approaches. In addition, our online testing results confirm the framework’s effectiveness in delivering superior recommendations and effectively reducing user churn by 4.95%.

Epistatic interactions of gene loci often determine complex trait phenotypes and may indicate the underlying molecular mechanisms of traits and diseases. Yet, the inference of epistatic interactions and gene–gene networks remains challenging. Neural networks have become successful in classifying complex data, revolutionizing various fields. However, their complexity does not reveal how they combine input features, and this lack of interpretability limits their use with genetic data. We thus introduce EpiDetect, a novel framework for discovering interactions between input features (single-nucleotide polymorphisms—SNPs—in our setting). EpiDetect neural-network–based classifiers detect interactions in systolic, diastolic, and pulse pressure genome-wide association datasets. Central to EpiDetect is EpiCID, a novel explainability algorithm for neural networks. Using EpiCID, we identified a network of highly interactive SNPs, performed centrality analysis to pinpoint central SNPs and genes, and outperformed established epistatic-interaction-detection algorithms. Moreover, pathway analysis uncovered well-known and novel pathways that could play a significant role in blood pressure traits, opening up new research directions.

Federated Learning (FL) enables collaborative training of Machine Learning (ML) models across decentralized clients while preserving data privacy. One of the challenges that FL faces is when the clients’ data is not independent and identically distributed (non-IID). It is, therefore, crucial to quantify how non-IID data impacts performance. However, due to the limited number of federated data available, it is not easy to carry out real-world simulations. In this work, we propose for the first time (1) the Hist-Dirichlet-based and Min-Size-Dirichlet methods for partitioning data into multiple nodes using the features and quantity distribution and the Dirichlet distribution. We use the (2) Jensen-Shannon and Hellinger distances for quantifying the degree of IID data. Moreover, we implemented (3) state-of-the-art partitioning methods based on the labels’ distribution across clients. All our proposals are open-source in a library called FedArtML, publicly available on PyPI. It facilitates research on cross-silo and cross-device FL, allowing a systematic and controlled partition of centralized datasets using the label, features, and quantity skewness. To demonstrate the value of our proposed methods and the robustness of FedArtML, we experimented in the ECG arrhythmia detection field with Physionet 2020 data. Our results demonstrate that our tool generates federated datasets for multi-client model training and accurately measures client distribution heterogeneity. Our approach achieves 48% higher non-IID-ness than existing feature skew methods, providing more granularity. Furthermore, we validate our simulated federated datasets against real-world data, revealing only a 2% F1-Score difference, affirming the method’s real-life applicability.

Can we assess a priori how well a knowledge graph embedding will perform on a specific downstream task and in a specific part of the knowledge graph? Knowledge graph embeddings (KGEs) represent entities (e.g., ''da Vinci,'' ''Mona Lisa'') and relationships (e.g., ''painted'') of a knowledge graph (KG) as vectors. KGEs are generated by optimizing an embedding score, which assesses whether a triple (e.g., ''da Vinci,'' "painted,'' ''Mona Lisa'') exists in the graph. KGEs have been proven effective in a variety of web-related downstream tasks, including, for instance, predicting relationship(s) among entities. However, the problem of anticipating the performance of a given KGE in a certain downstream task and locally to a specific individual triple, has not been tackled so far. In this paper, we fill this gap withReliK, a Reli ability measure for K GEs. ReliK relies solely on KGE embedding scores, is task- and KGE-agnostic, and requires no further KGE training. As such, it is particularly appealing for semantic web applications which call for testing multiple KGE methods on various parts of the KG and on each individual downstream task. Through extensive experiments, we attest thatReliK correlates well with both common downstream tasks, such as tail/relation prediction and triple classification, as well as advanced downstream tasks, such as rule mining and question answering, while preserving locality.