SIA: Symbolic Interpretability for Anticipatory Deep Reinforcement Learning in Network Control

2026-01-29

AI Summary

This work addresses the limited interpretability of existing prediction-augmented deep reinforcement learning (DRL) methods, which hinders validation of how predictions influence decision-making. To overcome this, the authors propose a real-time explainable framework that integrates symbolic AI with a KPI knowledge graph, enabling— for the first time—symbolic explanations of prediction-augmented DRL agents. The framework introduces an influence-scoring mechanism that facilitates online policy optimization without requiring retraining. It achieves sub-millisecond explanation latency, offering over 200× speedup compared to state-of-the-art explainable AI (XAI) approaches. Evaluated on video streaming and RAN slicing tasks, the method improves average bitrate by 9% and cumulative reward by 25%, respectively, demonstrating significant gains in policy performance.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Computer Vision: Visual Reasoning & Symbolic Representations
• Machine Learning: Transparent, Interpretable, Explainable ML

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Deep reinforcement learning (DRL) promises adaptive control for future mobile networks but conventional agents remain reactive: they act on past and current measurements and cannot leverage short-term forecasts of exogenous KPIs such as bandwidth. Augmenting agents with predictions can overcome this temporal myopia, yet uptake in networking is scarce because forecast-aware agents act as closed-boxes; operators cannot tell whether predictions guide decisions or justify the added complexity. We propose SIA, the first interpreter that exposes in real time how forecast-augmented DRL agents operate. SIA fuses Symbolic AI abstractions with per-KPI Knowledge Graphs to produce explanations, and includes a new Influence Score metric. SIA achieves sub-millisecond speed, over 200x faster than existing XAI methods. We evaluate SIA on three diverse networking use cases, uncovering hidden issues, including temporal misalignment in forecast integration and reward-design biases that trigger counter-productive policies. These insights enable targeted fixes: a redesigned agent achieves a 9% higher average bitrate in video streaming, and SIA's online Action-Refinement module improves RAN-slicing reward by 25% without retraining. By making anticipatory DRL transparent and tunable, SIA lowers the barrier to proactive control in next-generation mobile networks.

Problem

• Deep Reinforcement Learning
• Network Control
• Forecast-aware Agents
• Explainability
• Temporal Myopia

Innovation

• Symbolic AI
• Explainable AI (XAI)
• Anticipatory DRL
• Knowledge Graph
• Influence Score