-
Abstract: The stochastic inventory control problem under censored demands is a fundamental problem in revenue and supply chain management. A simple class of policies…
Learning in structured MDPs with convex cost…
From Oluwasanmi Koyejo
17 plays 0 -
Abstract: Policy gradient methods apply to complex, poorly understood, control problems by performing stochastic gradient descent over a parameterized class of…
On the Global Convergence and Approximation…
From Oluwasanmi Koyejo
9 plays 0 -
Abstract: Recent investigations into infinitely-wide deep neural networks have given rise to intriguing connections between deep networks, kernel methods, and Gaussian…
A Phase Transition in Gradient Descent for Wide,…
From Oluwasanmi Koyejo
31 plays 0 -
Abstract: Does deep learning actually need to be deep? In this talk, I will present some of our recent and ongoing work on Deep Equilibrium (DEQ) Models, an approach…
Equilibrium approaches to deep learning: One…
From Oluwasanmi Koyejo
38 plays 0 -
Abstract: We study the problem of recovering a planted matching in randomly weighted complete bipartite graphs with n nodes. For some unknown perfect matching M*, the…
Planted Matching Problem, Mehrdad Moharrami; IDS2…
From Oluwasanmi Koyejo
38 plays 0 -
Abstract: In numerous decision-making processes such as algorithm portfolio design and adaptive routing, each action incurs a random cost and yields a random and…
Budget-Constrained Learning and Optimization with…
From Oluwasanmi Koyejo
21 plays 0 -
Abstract: Modern Reinforcement Learning (RL) is commonly applied to practical problems with an enormous number of states, where function approximation such as deep…
Provably Efficient Exploration in Reinforcement…
From Oluwasanmi Koyejo
41 plays 0 -
Abstract: Flexible, yet interpretable, models for the second-order temporal structure are needed in scientific analyses of high-dimensional data. We develop a…
A Nonconvex Framework for Structured Dynamic…
From Oluwasanmi Koyejo
25 plays 0 -
Byzantine Tolerance for Distributed SGD; Cong Xie, IDS2 seminar series
Byzantine Tolerance for Distributed SGD; Cong…
From Oluwasanmi Koyejo
9 plays 0 -
Abstract: How does a 110-layer ResNet learn a high-complexity classifier using relatively few training examples and short training time? We present a theory towards…
Backward feature correction: How can deep…
From Oluwasanmi Koyejo
21 plays 0 -
Abstract: System identification has a long history with several well-established methods, in particular for learning linear dynamical systems from input/output data.…
A fresh look at a classical system identification…
From Oluwasanmi Koyejo
11 plays 0
Search for ""
Public, Restricted and Moderated