Matches in SemOpenAlex for { <https://semopenalex.org/work/W3179664243> ?p ?o ?g. }
- W3179664243 abstract "Spatial self-attention layers, in the form of Non-Local blocks, introduce long-range dependencies in Convolutional Neural Networks by computing pairwise similarities among all possible positions. Such pairwise functions underpin the effectiveness of non-local layers, but also determine a complexity that scales quadratically with respect to the input size both in space and time. This is a severely limiting factor that practically hinders the applicability of non-local blocks to even moderately sized inputs. Previous works focused on reducing the complexity by modifying the underlying matrix operations, however in this work we aim to retain full expressiveness of non-local layers while keeping complexity linear. We overcome the efficiency limitation of non-local blocks by framing them as special cases of 3rd order polynomial functions. This fact enables us to formulate novel fast Non-Local blocks, capable of reducing the complexity from quadratic to linear with no loss in performance, by replacing any direct computation of pairwise similarities with element-wise multiplications. The proposed method, which we dub as Poly-NL, is competitive with state-of-the-art performance across image recognition, instance segmentation, and face detection tasks, while having considerably less computational overhead." @default.
- W3179664243 created "2021-07-19" @default.
- W3179664243 creator A5004562735 @default.
- W3179664243 creator A5051613830 @default.
- W3179664243 creator A5068456312 @default.
- W3179664243 creator A5073110676 @default.
- W3179664243 creator A5080553022 @default.
- W3179664243 creator A5085009855 @default.
- W3179664243 date "2021-07-06" @default.
- W3179664243 modified "2023-09-27" @default.
- W3179664243 title "Poly-NL: Linear Complexity Non-local Layers with Polynomials." @default.
- W3179664243 cites W1861492603 @default.
- W3179664243 cites W2024165284 @default.
- W3179664243 cites W2056370875 @default.
- W3179664243 cites W2064675550 @default.
- W3179664243 cites W2097073572 @default.
- W3179664243 cites W2104657103 @default.
- W3179664243 cites W2117539524 @default.
- W3179664243 cites W2120190345 @default.
- W3179664243 cites W2136163184 @default.
- W3179664243 cites W2147880316 @default.
- W3179664243 cites W2170653751 @default.
- W3179664243 cites W2179571827 @default.
- W3179664243 cites W2194775991 @default.
- W3179664243 cites W2525246036 @default.
- W3179664243 cites W2556967412 @default.
- W3179664243 cites W2565639579 @default.
- W3179664243 cites W2622263826 @default.
- W3179664243 cites W2626778328 @default.
- W3179664243 cites W2883502031 @default.
- W3179664243 cites W2884585870 @default.
- W3179664243 cites W2937843571 @default.
- W3179664243 cites W2940744433 @default.
- W3179664243 cites W2941814890 @default.
- W3179664243 cites W2945164022 @default.
- W3179664243 cites W2950141105 @default.
- W3179664243 cites W2953484813 @default.
- W3179664243 cites W2954930822 @default.
- W3179664243 cites W2955058313 @default.
- W3179664243 cites W2962858109 @default.
- W3179664243 cites W2963091558 @default.
- W3179664243 cites W2963150697 @default.
- W3179664243 cites W2963266682 @default.
- W3179664243 cites W2963319519 @default.
- W3179664243 cites W2963351448 @default.
- W3179664243 cites W2963420686 @default.
- W3179664243 cites W2963495494 @default.
- W3179664243 cites W2963566548 @default.
- W3179664243 cites W2963807318 @default.
- W3179664243 cites W2965669158 @default.
- W3179664243 cites W2981413347 @default.
- W3179664243 cites W2982220924 @default.
- W3179664243 cites W2994821360 @default.
- W3179664243 cites W2997747012 @default.
- W3179664243 cites W3021892279 @default.
- W3179664243 cites W3023064088 @default.
- W3179664243 cites W3033529678 @default.
- W3179664243 cites W3034217077 @default.
- W3179664243 cites W3034345703 @default.
- W3179664243 cites W3034540350 @default.
- W3179664243 cites W3034573343 @default.
- W3179664243 cites W3085139254 @default.
- W3179664243 cites W3120633509 @default.
- W3179664243 cites W3125056032 @default.
- W3179664243 cites W7746136 @default.
- W3179664243 cites W78159342 @default.
- W3179664243 hasPublicationYear "2021" @default.
- W3179664243 type Work @default.
- W3179664243 sameAs 3179664243 @default.
- W3179664243 citedByCount "0" @default.
- W3179664243 crossrefType "posted-content" @default.
- W3179664243 hasAuthorship W3179664243A5004562735 @default.
- W3179664243 hasAuthorship W3179664243A5051613830 @default.
- W3179664243 hasAuthorship W3179664243A5068456312 @default.
- W3179664243 hasAuthorship W3179664243A5073110676 @default.
- W3179664243 hasAuthorship W3179664243A5080553022 @default.
- W3179664243 hasAuthorship W3179664243A5085009855 @default.
- W3179664243 hasConcept C11413529 @default.
- W3179664243 hasConcept C129844170 @default.
- W3179664243 hasConcept C154945302 @default.
- W3179664243 hasConcept C179799912 @default.
- W3179664243 hasConcept C184898388 @default.
- W3179664243 hasConcept C195956108 @default.
- W3179664243 hasConcept C2524010 @default.
- W3179664243 hasConcept C311688 @default.
- W3179664243 hasConcept C33923547 @default.
- W3179664243 hasConcept C41008148 @default.
- W3179664243 hasConcept C45374587 @default.
- W3179664243 hasConcept C80444323 @default.
- W3179664243 hasConcept C81363708 @default.
- W3179664243 hasConceptScore W3179664243C11413529 @default.
- W3179664243 hasConceptScore W3179664243C129844170 @default.
- W3179664243 hasConceptScore W3179664243C154945302 @default.
- W3179664243 hasConceptScore W3179664243C179799912 @default.
- W3179664243 hasConceptScore W3179664243C184898388 @default.
- W3179664243 hasConceptScore W3179664243C195956108 @default.
- W3179664243 hasConceptScore W3179664243C2524010 @default.
- W3179664243 hasConceptScore W3179664243C311688 @default.
- W3179664243 hasConceptScore W3179664243C33923547 @default.
- W3179664243 hasConceptScore W3179664243C41008148 @default.