On the Global Optimality of Model-Agnostic Meta-Learning
원문 초록
Model-agnostic meta-learning (MAML) formulates meta-learning as a bilevel optimization problem, where the inner level solves each subtask based on a shared prior, while the outer level searches for the optimal shared prior by optimizing its aggregated performance over all the subtasks. Despite its empirical success, MAML remains less understood in theory, especially in terms of its global optimality, due to the nonconvexity of the meta-objective (the outer-level objective). To bridge such a gap between theory and practice, we characterize the optimality gap of the stationary points attained by MAML for both reinforcement learning and supervised learning, where the inner-level and outer-level problems are solved via first-order optimization methods. In particular, our characterization connects the optimality gap of such stationary points with (i) the functional geometry of inner-level objectives and (ii) the representation power of function approximators, including linear models and neural networks. To the best of our knowledge, our analysis establishes the global optimality of MAML with nonconvex meta-objectives for the first time.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Differentially Private Bilevel Optimization: Efficient Algorithms with Near-Optimal Rates
Efficient Curvature-Aware Hypergradient Approximation for Bilevel Optimization
On the Convergence of Adam-Type Algorithm for Bilevel Optimization under Unbounded Smoothness
qNBO: quasi-Newton Meets Bilevel Optimization
Evaluating Data Influence in Meta Learning
Targeted Data-driven Regularization for Out-of-Distribution Generalization
토론
이 논문과 연결된 포럼 글이 아직 없습니다.