A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Filters
Fast Distributionally Robust Learning with Variance Reduced Min-Max Optimization
[article]
2022
arXiv
pre-print
Preserved Fulltext
Other Versions
Key to our results is the use of variance reduction and random reshuffling to accelerate stochastic min-max optimization, the analysis of which may be of independent interest. ...
Distributionally robust supervised learning (DRSL) is emerging as a key paradigm for building reliable machine learning systems for real-world applications -- reflecting the need for classifiers and predictive ...
Conclusions We studied Wasserstein distributionally robust supervised learning through the lens of min-max optimization. ...
arXiv:2104.13326v2
fatcat:bqobb3ieijajnhykasqjd6masa
Group Distributionally Robust Reinforcement Learning with Hierarchical Latent Variables
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Robust RL has been applied to deal with task ambiguity, but may result in over-conservative policies. ...
To balance the worst-case (robustness) and average performance, we propose Group Distributionally Robust Markov Decision Process (GDR-MDP), a flexible hierarchical MDP formulation that encodes task groups ...
The Group Distributionally Robust Bellman optimality equation is V π t (b t , s t ) = max πt min bt∈Cb t E bt(z) E µz(m) E πt E Rm [r t ] + γ st+1 P m (s t+1 |s t , a t )V π t+1 (b t+1 , s t+1 ) . ...
arXiv:2210.12262v1
fatcat:sruqwgyeefeb3pw2xwe5ygfxri
Compositional federated learning: Applications in distributionally robust averaging and meta learning
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
data mining and machine learning problems with a hierarchical structure such as distributionally robust FL and model-agnostic meta learning (MAML). ...
In particular, we first transform the distributionally robust FL (i.e., a minimax optimization problem) into a simple composition optimization problem by using KL divergence regularization. ...
Index Terms-Federated Learning, Composition Optimization, Distributionally Robust, Meta Learning, Model Agnostic. ...
arXiv:2106.11264v3
fatcat:anekjuzz65en7etnlufbeyncuy
Generalized Wasserstein Dice Score, Distributionally Robust Deep Learning, and Ranger for brain tumor segmentation: BraTS 2020 challenge
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
, corresponding to distributionally robust optimization, and a non-standard optimizer, Ranger. ...
Distributionally robust optimization is a generalization of empirical risk minimization that accounts for the presence of underrepresented subdomains in the training dataset. ...
Changing the Optimizer: Ranger RAdam [23] is a modification of the Adam optimizer [21] that aims at reducing the variance of the adaptive learning rate of Adam in the early-stage of training. ...
arXiv:2011.01614v2
fatcat:sns55rjuordated7d3xlxpv6je
Distributionally Robust Surrogate Optimal Control for High-Dimensional Systems
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
This work is motivated by the ongoing challenges of safety, computation, and optimality in high-dimensional optimal control. We address these key questions with the following approach. ...
This paper presents a novel methodology for tractably solving optimal control and offline reinforcement learning problems for high-dimensional systems. ...
This highlights a comparative Figure 4 shows the optimal fast charging results for versions of our algorithm with and without distributionally robust optimization. ...
arXiv:2105.10070v2
fatcat:cx5bttn4afgitoh77juq4z4any
Reliable Off-policy Evaluation for Reinforcement Learning
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Leveraging methodologies from distributionally robust optimization, we show that with proper selection of the size of the distributional uncertainty set, these estimates serve as confidence bounds with ...
Our results are also generalized to batch reinforcement learning and are supported by empirical analysis. ...
Smirnova E, Dohmatob E, Mary J ( ) Distributionally robust reinforcement learning. ArXiv abs/ . . Song J, Zhao C ( ) Optimistic distributionally robust policy optimization. arXiv preprint arXiv: . . ...
arXiv:2011.04102v2
fatcat:hd4mbmxxtjd65iprydrf2xygxy
An approach to the distributionally robust shortest path problem
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Under some additional assumptions the resulting distributionally robust shortest path problem (DRSPP) admits equivalent robust and mixed-integer programming (MIP) reformulations. ...
The ambiguity set is formed by all distributions that satisfy prescribed linear first-order moment constraints with respect to subsets of arcs and individual probability constraints with respect to particular ...
[37] for the distributionally robust max-min problems with a marginal moment ambiguity set. ...
arXiv:1910.08744v3
fatcat:ewhvgxddxbgwdd3hten37rk4fe
DR-DSGD: A Distributionally Robust Decentralized Learning Algorithm over Graphs
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
By adding a Kullback-Liebler regularization function to the robust min-max optimization problem, the learning problem can be reduced to a modified robust minimization problem and solved efficiently. ...
In this paper, we propose to solve a regularized distributionally robust learning problem in the decentralized setting, taking into account the data distribution shift. ...
In this case, the distributionally robust empirical loss problem is given by the following min-max optimization problem min Θ∈R d max λ∈∆ K i=1 λ i f i (Θ). ( 5 ) Although several distributed algorithms ...
arXiv:2208.13810v2
fatcat:nik3iefajjb2xcniw3lrbksxqq
Distributionally Robust Optimization: A review on theory and applications
2022
Numerical Algebra, Control and Optimization
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
<p style='text-indent:20px;'>In this paper, we survey the primary research on the theory and applications of distributionally robust optimization (DRO). ...
We start with reviewing the modeling power and computational attractiveness of DRO approaches, induced by the ambiguity sets structure and tractable robust counterpart reformulations. ...
[66] also proposed their work in distributionally robust learning with unlabeled data. ...
doi:10.3934/naco.2021057
fatcat:cluvjzkrhfbc7o2446fhjrmbji
Efficient Distributionally Robust Bayesian Optimization with Worst-case Sensitivity
2022
International Conference on Machine Learning
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
In distributionally robust Bayesian optimization (DRBO), an exact computation of the worst-case expected value requires solving an expensive convex optimization problem. ...
We provide a regret bound for our novel DRBO algorithm with the fast approximation, and empirically show it is competitive with that using the exact worst-case expected value while incurring significantly ...
Distributionally robust BO (DRBO) was concurrently introduced by Kirschner et al. (2020) and Nguyen et al. (2020) and adapts the framework of distributionally robust optimization (DRO) in operations ...
dblp:conf/icml/TayFDLL22
fatcat:s2qsht2wwnhojouet2zmwe4tye
Efficient Stochastic Gradient Descent for Learning with Distributionally Robust Optimization
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2018 pre-print arXiv:1805.08728v1 fatcat:45tgt4uafjhqpgihchrliqikyu
Distributionally robust optimization (DRO) problems are increasingly seen as a viable method to train machine learning models for improved model generalization. ...
These min-max formulations, however, are more difficult to solve. We therefore provide a new stochastic gradient descent algorithm to efficiently solve this DRO formulation. ...
Distributionally Robust Learning Consider a general formulation of the distributionally robust optimization problem of active interest. ...
arXiv:1805.08728v2
fatcat:u2jeqbumtbbfjlvelv2gt7nicy
Wasserstein Distributionally Robust Inverse Multiobjective Optimization
2021
PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
To hedge against the uncertainties in the hypothetical DMP, the data, and the parameter space, we investigate in this paper the distributionally robust approach for inverse multiobjective optimization. ...
We then formulate a Wasserstein distributionally robust inverse multiobjective optimization problem (WRO-IMOP) that minimizes a worst-case expected loss function, where the worst case is taken over all ...
Algorithm 1 Wasserstein Distributionally Robust IMOP if CV i > 0 then let Y i ← Y i ∪ { y i } end if 8: end for 9: until max i∈[N ] CV i ≤ δ 10: Output: a δ-optimal solution θ N of (3) Using the above ...
doi:10.1609/aaai.v35i7.16739
fatcat:vugh7jk5hnalzf7l525z4eau3q
Distributionally Robust Optimization
[chapter]
2018
Optimization Algorithms - Examples
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
This chapter presents a class of distributionally robust optimization problems in which a decision-maker has to choose an action in an uncertain environment. ...
This leads to a distributionally robust optimization problem. Simple algorithms, whose dynamics are inspired from the gradient flows, are proposed to find local optima. ...
The presented methodology is simple and reduces significantly the dimensionality of the distributionally robust optimization. ...
doi:10.5772/intechopen.76686
fatcat:3s5nhfcjwncsjfxqp6mlnkdnhe
Sinkhorn Distributionally Robust Optimization
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We study distributionally robust optimization (DRO) with Sinkhorn distance – a variant of Wasserstein distance based on entropic regularization. ...
We demonstrate that our proposed algorithm finds δ-optimal solution of the new DRO formulation with computation cost Õ(δ^-3) and memory cost Õ(δ^-2), and the computation cost further improves to Õ(δ^-2 ...
IEEE Transactions on Automatic Control ( ): -. [ ] Yu Y, Lin T, Mazumdar EV, Jordan M ( ) Fast distributionally robust learning with variance-reduced min-max optimization. ...
arXiv:2109.11926v2
fatcat:jkoyh2rrqvdvxfdke75k7liv54
Gradient Descent Ascent for Minimax Problems on Riemannian Manifolds
[article]
2023
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf.
Other Versions
To further reduce the sample complexity, we propose an accelerated Riemannian stochastic gradient descent ascent (Acc-RSGDA) algorithm based on the momentum-based variance-reduced technique. ...
Extensive experimental results on the robust distributional optimization and robust Deep Neural Networks (DNNs) training over Stiefel manifold demonstrate efficiency of our algorithms. ...
Subsequently, [43] presented fast stochastic variance-reduced methods to Riemannian manifold optimization. ...
arXiv:2010.06097v5
fatcat:tgz3wv7abzaqbnphfwt43j7hoa
« Previous
Showing results 1 — 15 out of 426 results