Matches in SemOpenAlex for { <https://semopenalex.org/work/W2592457170> ?p ?o ?g. }
- W2592457170 abstract "This paper presents OptNet, a network architecture that integrates optimization problems (here, specifically in the form of quadratic programs) as individual layers in larger end-to-end trainable deep networks. These layers encode constraints and complex dependencies between the hidden states that traditional convolutional and fully-connected layers often cannot capture. In this paper, we explore the foundations for such an architecture: we show how techniques from sensitivity analysis, bilevel optimization, and implicit differentiation can be used to exactly differentiate through these layers and with respect to layer parameters; we develop a highly efficient solver for these layers that exploits fast GPU-based batch solves within a primal-dual interior point method, and which provides backpropagation gradients with virtually no additional cost on top of the solve; and we highlight the application of these approaches in several problems. In one notable example, we show that the method is capable of learning to play mini-Sudoku (4x4) given just input and output games, with no a priori information about the rules of the game; this highlights the ability of our architecture to learn hard constraints better than other neural architectures." @default.
- W2592457170 created "2017-03-16" @default.
- W2592457170 creator A5030666794 @default.
- W2592457170 creator A5075035644 @default.
- W2592457170 date "2017-03-01" @default.
- W2592457170 modified "2023-09-27" @default.
- W2592457170 title "OptNet: Differentiable Optimization as a Layer in Neural Networks" @default.
- W2592457170 cites W1522301498 @default.
- W2592457170 cites W1569098853 @default.
- W2592457170 cites W1585773866 @default.
- W2592457170 cites W1601741115 @default.
- W2592457170 cites W1978259121 @default.
- W2592457170 cites W2000985550 @default.
- W2592457170 cites W2016211524 @default.
- W2592457170 cites W2021468336 @default.
- W2592457170 cites W2047993072 @default.
- W2592457170 cites W2097113878 @default.
- W2592457170 cites W2098617596 @default.
- W2592457170 cites W2117402460 @default.
- W2592457170 cites W2148862211 @default.
- W2592457170 cites W2163112044 @default.
- W2592457170 cites W2186629860 @default.
- W2592457170 cites W2187061624 @default.
- W2592457170 cites W2464234964 @default.
- W2592457170 cites W2505728881 @default.
- W2592457170 cites W2599275287 @default.
- W2592457170 cites W2798766386 @default.
- W2592457170 cites W2953246223 @default.
- W2592457170 cites W2963266548 @default.
- W2592457170 cites W2964205912 @default.
- W2592457170 hasPublicationYear "2017" @default.
- W2592457170 type Work @default.
- W2592457170 sameAs 2592457170 @default.
- W2592457170 citedByCount "40" @default.
- W2592457170 countsByYear W25924571702017 @default.
- W2592457170 countsByYear W25924571702018 @default.
- W2592457170 countsByYear W25924571702019 @default.
- W2592457170 countsByYear W25924571702020 @default.
- W2592457170 countsByYear W25924571702021 @default.
- W2592457170 crossrefType "posted-content" @default.
- W2592457170 hasAuthorship W2592457170A5030666794 @default.
- W2592457170 hasAuthorship W2592457170A5075035644 @default.
- W2592457170 hasConcept C104317684 @default.
- W2592457170 hasConcept C108583219 @default.
- W2592457170 hasConcept C111472728 @default.
- W2592457170 hasConcept C11413529 @default.
- W2592457170 hasConcept C127413603 @default.
- W2592457170 hasConcept C134306372 @default.
- W2592457170 hasConcept C137836250 @default.
- W2592457170 hasConcept C138885662 @default.
- W2592457170 hasConcept C154945302 @default.
- W2592457170 hasConcept C155032097 @default.
- W2592457170 hasConcept C165696696 @default.
- W2592457170 hasConcept C178790620 @default.
- W2592457170 hasConcept C185592680 @default.
- W2592457170 hasConcept C199360897 @default.
- W2592457170 hasConcept C202615002 @default.
- W2592457170 hasConcept C21200559 @default.
- W2592457170 hasConcept C24326235 @default.
- W2592457170 hasConcept C2524010 @default.
- W2592457170 hasConcept C2778770139 @default.
- W2592457170 hasConcept C2779227376 @default.
- W2592457170 hasConcept C28719098 @default.
- W2592457170 hasConcept C33923547 @default.
- W2592457170 hasConcept C38652104 @default.
- W2592457170 hasConcept C41008148 @default.
- W2592457170 hasConcept C50644808 @default.
- W2592457170 hasConcept C55493867 @default.
- W2592457170 hasConcept C66746571 @default.
- W2592457170 hasConcept C75553542 @default.
- W2592457170 hasConcept C81363708 @default.
- W2592457170 hasConceptScore W2592457170C104317684 @default.
- W2592457170 hasConceptScore W2592457170C108583219 @default.
- W2592457170 hasConceptScore W2592457170C111472728 @default.
- W2592457170 hasConceptScore W2592457170C11413529 @default.
- W2592457170 hasConceptScore W2592457170C127413603 @default.
- W2592457170 hasConceptScore W2592457170C134306372 @default.
- W2592457170 hasConceptScore W2592457170C137836250 @default.
- W2592457170 hasConceptScore W2592457170C138885662 @default.
- W2592457170 hasConceptScore W2592457170C154945302 @default.
- W2592457170 hasConceptScore W2592457170C155032097 @default.
- W2592457170 hasConceptScore W2592457170C165696696 @default.
- W2592457170 hasConceptScore W2592457170C178790620 @default.
- W2592457170 hasConceptScore W2592457170C185592680 @default.
- W2592457170 hasConceptScore W2592457170C199360897 @default.
- W2592457170 hasConceptScore W2592457170C202615002 @default.
- W2592457170 hasConceptScore W2592457170C21200559 @default.
- W2592457170 hasConceptScore W2592457170C24326235 @default.
- W2592457170 hasConceptScore W2592457170C2524010 @default.
- W2592457170 hasConceptScore W2592457170C2778770139 @default.
- W2592457170 hasConceptScore W2592457170C2779227376 @default.
- W2592457170 hasConceptScore W2592457170C28719098 @default.
- W2592457170 hasConceptScore W2592457170C33923547 @default.
- W2592457170 hasConceptScore W2592457170C38652104 @default.
- W2592457170 hasConceptScore W2592457170C41008148 @default.
- W2592457170 hasConceptScore W2592457170C50644808 @default.
- W2592457170 hasConceptScore W2592457170C55493867 @default.
- W2592457170 hasConceptScore W2592457170C66746571 @default.
- W2592457170 hasConceptScore W2592457170C75553542 @default.
- W2592457170 hasConceptScore W2592457170C81363708 @default.