Temporal-difference Variational Continual Learning

Abstract

Machine Learning models in real-world applications must continuously learn new tasks to adapt to shifts in the data-generating distribution. Yet, for Continual Learning (CL), models often struggle to balance learning new tasks (plasticity) with retaining previous knowledge (memory stability). Consequently, they are susceptible to Catastrophic Forgetting, which degrades performance and undermines the reliability of deployed systems. In the Bayesian CL literature, variational methods tackle this challenge by employing a learning objective that recursively updates the posterior distribution while constraining it to stay close to its previous estimate. Nonetheless, we argue that these methods may be ineffective due to compounding approximation errors over successive recursions. To mitigate this, we propose new learning objectives that integrate the regularization effects of multiple previous posterior estimations, preventing individual errors from dominating future posterior updates and co

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keymelo2024temporal

Related papers

• Prediction And Control In Continual Reinforcement Learning (2023)0.00
• Beyond Supervised Continual Learning: A Review (2022)0.00
• Continual Visual Reinforcement Learning With A Life-long World Model (2023)2.26
• The Hierarchical Adaptive Forgetting Variational Filter (2018)0.00
• Task-agnostic Continual Reinforcement Learning: Gaining Insights And Overcoming Challenges (2022)0.00
• Simple Recipe Works: Vision-language-action Models Are Natural Continual Learners With Reinforcement Learning (2026)0.00
• Continual Learning Using World Models For Pseudo-rehearsal (2019)0.00
• Discerning Temporal Difference Learning (2023)0.00