Explainable Recommendations in Large-Scale Content Feeds: Data Structures and Algorithms for Real-Time Reasoning Labels

Lev Fedorov

Citation: Lev Fedorov, "Explainable Recommendations in Large-Scale Content Feeds: Data Structures and Algorithms for Real-Time Reasoning Labels", Universal Library of Innovative Research and Studies, Volume 03, Issue 01.

Copyright: This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Under conditions of exponential growth in the volume of digital content and increasing complexity of machine learning algorithms, recommender systems (RS) have transformed from auxiliary navigation tools into a critical infrastructure of the digital economy. However, the dominance of deep neural networks (DNN) and large language models (LLM) has led to the black box problem, where the opacity of decision-making undermines user trust and conflicts with new regulatory frameworks such as the EU AI Act. This work addresses the fundamental trade-off between recommendation accuracy and interpretability under strict real-time constraints (<100 ms). The study introduces a new architecture, Neuro-Symbolic Reasoning Label (NSRL), which employs a hybrid neuro-symbolic approach to precompute causal reasoning paths on knowledge graphs and encode them into compact data structures called Reasoning Labels. Experimental results on large-scale datasets (Amazon-Book, Yelp2018) show that the proposed method achieves ranking accuracy metrics (NDCG@20 ~0.0815) comparable to state-of-the-art models (SASRec, KGAT), while providing high explanation fidelity (fidelity > 0.8) and inference latency of 45 ms. The study lays the theoretical and practical foundation for building reliable, agent-based next-generation recommender systems.


Keywords: Recommender Systems, Explainable Artificial Intelligence (XAI), Knowledge Graphs, Neuro-Symbolic AI, EU AI Act, Reasoning Labels, Real-Time Inference.

Download doi https://doi.org/10.70315/uloap.ulirs.2026.0301017