AAAI 2021 accepted papers on Optimization
42 of the 1,654 papers accepted at AAAI 2021 (2.5%) are on Optimization, compared with 1.7% at AAAI 2020. They fall into 4 topics.
See AAAI 2021 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest papers by meaning and how crowded the spot has become.
Topics and papers
Optimizers · 22 papers
- ADAHESSIAN: An Adaptive Second Order Optimizer for Machine Learning (190 citations)
- Sample Efficient Reinforcement Learning with REINFORCE (103 citations)
- On the Adequacy of Untuned Warmup for Adaptive Optimization (34 citations)
- Frugal Optimization for Cost-related Hyperparameters (25 citations)
- On the Verification of Neural ODEs with Stochastic Guarantees (17 citations)
- Large Batch Optimization for Deep Learning Using New Complete Layer-Wise Adaptive Rate Scaling (14 citations)
- Differentially Private Stochastic Coordinate Descent (10 citations)
- Explaining Neural Matrix Factorization with Gradient Rollback (9 citations)
- A Trace-restricted Kronecker-Factored Approximation to Natural Gradient (7 citations)
- THOR, Trace-based Hardware-driven Layer-Oriented Natural Gradient Descent Computation (7 citations)
- Physarum Powered Differentiable Linear Programming Layers and Applications (5 citations)
- Deterministic Mini-batch Sequencing for Training Deep Neural Networks (4 citations)
and 10 more in the interactive map.
Implicit bias and generalization · 8 papers
- Deep Frequency Principle Towards Understanding Why Deeper Learning Is Faster (44 citations)
- Stability and Generalization of Decentralized Stochastic Gradient Descent (25 citations)
- Adaptive Knowledge Driven Regularization for Deep Neural Networks (16 citations)
- A Deeper Look at the Hessian Eigenspectrum of Deep Neural Networks and its Applications to Regularization (12 citations)
- A Recipe for Global Convergence Guarantee in Deep Neural Networks (8 citations)
- Fine-grained Generalization Analysis of Vector-Valued Learning (3 citations)
- Characterizing the Loss Landscape in Non-Negative Matrix Factorization (1 citation)
- Why Adversarial Interaction Creates Non-Homogeneous Patterns: A Pseudo-Reaction-Diffusion Model for Turing Instability (1 citation)
Nonconvex and smooth optimization · 6 papers
- Increasing Iterate Averaging for Solving Saddle-Point Problems (6 citations)
- Adaptive Gradient Methods for Constrained Convex Optimization and Variational Inequalities (5 citations)
- Enhancing Parameter-Free Frank Wolfe with an Extra Subproblem (3 citations)
- On Convergence of Gradient Expected Sarsa(λ) (2 citations)
- Smooth Convex Optimization Using Sub-Zeroth-Order Oracles (2 citations)
- Gradient Descent Averaging and Primal-dual Averaging for Strongly Convex Optimization (1 citation)
Distributed and compressed · 6 papers
- Differentially Private and Communication Efficient Collaborative Learning (22 citations)
- Communication-Efficient Frank-Wolfe Algorithm for Nonconvex Decentralized Distributed Learning (12 citations)
- Step-Ahead Error Feedback for Distributed Training with Compressed Gradient (12 citations)
- Elastic Consistency: A Practical Consistency Model for Distributed Stochastic Gradient Descent (7 citations)
- STL-SGD: Speeding Up Local SGD with Stagewise Communication Period (7 citations)
- Asynchronous Optimization Methods for Efficient Training of Deep Neural Networks with Guarantees (3 citations)
