Abstract Reinforcement learning (RL) agents in sequential decision-making face several fundamental dilemmas or "paradoxes" that hinder real-world deployment. This paper provides a comprehensive analysis of paradox-aware reinforcement learning in the context of closed-loop time series systems. We focus on key theoretical challenges – such as the exploration-exploitation dilemma, the temporal credit assignment problem, the simulation-to-reality gap, and distributional shift – and examine how thes...

Abstract: Artificial intelligence (AI) systems must balance memory retention with the ability to forget irrelevant or harmful information. This paper provides a comprehensive overview of how memory-loss events – instances of forgetting or suppressing stored information – can be modeled in AI through specialized data structures and algorithms. We integrate theoretical and applied perspectives from deep learning (e.g. catastrophic forgetting in neural networks), transformer architectures (e.g. att...

Introduction Time travel to the past traditionally raises famous logical paradoxes in physics and fiction, the most well-known being the “grandfather paradox.” This paradox asks: what happens if a time traveler goes back and prevents their own grandparents from meeting, thus preventing the time traveler’s birth? Such a self-contradictory causal loop threatens the consistency of reality. Stephen Hawking humorously illustrated the issue by hosting a party for time travelers, sending the invitatio...

Abstract Artificial intelligence (AI) models that make sequential or time-dependent predictions often struggle with temporal consistency – the requirement that outputs vary smoothly and logically over time without erratic jumps. Abrupt or incoherent changes in predictions can degrade performance in applications from video analysis to time-series forecasting. This paper surveys recent advances in Causal Graph Neural Networks (GNNs) as a promising approach to enforce temporal consistency in AI pr...