FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048443606> ?p ?o ?g. }

• W3048443606 abstract "Multi-task learning aims at solving multiple machine learning tasks at the same time. A good solution to a multi-task learning problem should be generalizable in addition to being Pareto optimal. In this paper, we provide some insights on understanding the trade-off between Pareto efficiency and generalization as a result of parameterization in multi-task deep learning models. As a multi-objective optimization problem, enough parameterization is needed for handling task conflicts in a constrained solution space; however, from a multi-task generalization perspective, over-parameterization undermines the benefit of learning a shared representation which helps harder tasks or tasks with limited training examples. A delicate balance between multi-task generalization and multi-objective optimization is therefore needed for finding a better trade-off between efficiency and generalization. To this end, we propose a method of under-parameterized self-auxiliaries for multi-task models to achieve the best of both worlds. It is task-agnostic and works with other multi-task learning algorithms. Empirical results show that small towers of under-parameterized self-auxiliaries can make big differences in improving Pareto efficiency in various multi-task applications." @default.
• W3048443606 created "2020-08-18" @default.
• W3048443606 creator A5015729655 @default.
• W3048443606 creator A5021315343 @default.
• W3048443606 creator A5028125399 @default.
• W3048443606 creator A5037954928 @default.
• W3048443606 creator A5050182143 @default.
• W3048443606 creator A5079085366 @default.
• W3048443606 creator A5086484914 @default.
• W3048443606 date "2020-08-13" @default.
• W3048443606 modified "2023-09-25" @default.
• W3048443606 title "Small Towers Make Big Differences" @default.
• W3048443606 cites W1536680647 @default.
• W3048443606 cites W1821462560 @default.
• W3048443606 cites W1882958252 @default.
• W3048443606 cites W1896424170 @default.
• W3048443606 cites W1970031898 @default.
• W3048443606 cites W1977310087 @default.
• W3048443606 cites W2008499862 @default.
• W3048443606 cites W2024608894 @default.
• W3048443606 cites W2060727197 @default.
• W3048443606 cites W2112796928 @default.
• W3048443606 cites W2117130368 @default.
• W3048443606 cites W2117482391 @default.
• W3048443606 cites W2140066605 @default.
• W3048443606 cites W2219888463 @default.
• W3048443606 cites W2295072214 @default.
• W3048443606 cites W2512971201 @default.
• W3048443606 cites W2556468274 @default.
• W3048443606 cites W2604763608 @default.
• W3048443606 cites W2609130030 @default.
• W3048443606 cites W2613718673 @default.
• W3048443606 cites W2624871570 @default.
• W3048443606 cites W2740941491 @default.
• W3048443606 cites W2750384547 @default.
• W3048443606 cites W2767434619 @default.
• W3048443606 cites W2787017828 @default.
• W3048443606 cites W2794356495 @default.
• W3048443606 cites W2802620382 @default.
• W3048443606 cites W2809290718 @default.
• W3048443606 cites W2890538051 @default.
• W3048443606 cites W2903852246 @default.
• W3048443606 cites W2913340405 @default.
• W3048443606 cites W2927589347 @default.
• W3048443606 cites W2937297214 @default.
• W3048443606 cites W2946233749 @default.
• W3048443606 cites W2950220847 @default.
• W3048443606 cites W2963518130 @default.
• W3048443606 cites W2963677766 @default.
• W3048443606 cites W2963877604 @default.
• W3048443606 cites W2970470314 @default.
• W3048443606 cites W2973172293 @default.
• W3048443606 cites W2995256665 @default.
• W3048443606 cites W2997359900 @default.
• W3048443606 cites W3208857029 @default.
• W3048443606 cites W633916340 @default.
• W3048443606 hasPublicationYear "2020" @default.
• W3048443606 type Work @default.
• W3048443606 sameAs 3048443606 @default.
• W3048443606 citedByCount "3" @default.
• W3048443606 countsByYear W30484436062021 @default.
• W3048443606 crossrefType "posted-content" @default.
• W3048443606 hasAuthorship W3048443606A5015729655 @default.
• W3048443606 hasAuthorship W3048443606A5021315343 @default.
• W3048443606 hasAuthorship W3048443606A5028125399 @default.
• W3048443606 hasAuthorship W3048443606A5037954928 @default.
• W3048443606 hasAuthorship W3048443606A5050182143 @default.
• W3048443606 hasAuthorship W3048443606A5079085366 @default.
• W3048443606 hasAuthorship W3048443606A5086484914 @default.
• W3048443606 hasConcept C11413529 @default.
• W3048443606 hasConcept C119857082 @default.
• W3048443606 hasConcept C126255220 @default.
• W3048443606 hasConcept C12713177 @default.
• W3048443606 hasConcept C127413603 @default.
• W3048443606 hasConcept C134306372 @default.
• W3048443606 hasConcept C137635306 @default.
• W3048443606 hasConcept C154945302 @default.
• W3048443606 hasConcept C165464430 @default.
• W3048443606 hasConcept C177148314 @default.
• W3048443606 hasConcept C17744445 @default.
• W3048443606 hasConcept C199539241 @default.
• W3048443606 hasConcept C201995342 @default.
• W3048443606 hasConcept C2776359362 @default.
• W3048443606 hasConcept C2778599509 @default.
• W3048443606 hasConcept C2780451532 @default.
• W3048443606 hasConcept C28006648 @default.
• W3048443606 hasConcept C33923547 @default.
• W3048443606 hasConcept C41008148 @default.
• W3048443606 hasConcept C94625758 @default.
• W3048443606 hasConceptScore W3048443606C11413529 @default.
• W3048443606 hasConceptScore W3048443606C119857082 @default.
• W3048443606 hasConceptScore W3048443606C126255220 @default.
• W3048443606 hasConceptScore W3048443606C12713177 @default.
• W3048443606 hasConceptScore W3048443606C127413603 @default.
• W3048443606 hasConceptScore W3048443606C134306372 @default.
• W3048443606 hasConceptScore W3048443606C137635306 @default.
• W3048443606 hasConceptScore W3048443606C154945302 @default.
• W3048443606 hasConceptScore W3048443606C165464430 @default.
• W3048443606 hasConceptScore W3048443606C177148314 @default.
• W3048443606 hasConceptScore W3048443606C17744445 @default.

• next