Matches in SemOpenAlex for { <https://semopenalex.org/work/W3048593389> ?p ?o ?g. }
- W3048593389 abstract "We propose $textit{Mish}$, a novel self-regularized non-monotonic activation function which can be mathematically defined as: $f(x)=xtanh(softplus(x))$. As activation functions play a crucial role in the performance and training dynamics in neural networks, we validated experimentally on several well-known benchmarks against the best combinations of architectures and activation functions. We also observe that data augmentation techniques have a favorable effect on benchmarks like ImageNet-1k and MS-COCO across multiple architectures. For example, Mish outperformed Leaky ReLU on YOLOv4 with a CSP-DarkNet-53 backbone on average precision ($AP_{50}^{val}$) by 2.1$%$ in MS-COCO object detection and ReLU on ResNet-50 on ImageNet-1k in Top-1 accuracy by $approx$1$%$ while keeping all other network parameters and hyperparameters constant. Furthermore, we explore the mathematical formulation of Mish in relation with the Swish family of functions and propose an intuitive understanding on how the first derivative behavior may be acting as a regularizer helping the optimization of deep neural networks. Code is publicly available at this https URL." @default.
- W3048593389 created "2020-08-18" @default.
- W3048593389 creator A5035144667 @default.
- W3048593389 date "2019-08-23" @default.
- W3048593389 modified "2023-10-01" @default.
- W3048593389 title "Mish: A Self Regularized Non-Monotonic Activation Function" @default.
- W3048593389 cites W104184427 @default.
- W3048593389 cites W114517082 @default.
- W3048593389 cites W1522301498 @default.
- W3048593389 cites W1533861849 @default.
- W3048593389 cites W1572063013 @default.
- W3048593389 cites W1665214252 @default.
- W3048593389 cites W1677182931 @default.
- W3048593389 cites W1861492603 @default.
- W3048593389 cites W1921523184 @default.
- W3048593389 cites W2064616889 @default.
- W3048593389 cites W2097117768 @default.
- W3048593389 cites W2102605133 @default.
- W3048593389 cites W2108598243 @default.
- W3048593389 cites W2109255472 @default.
- W3048593389 cites W2176412452 @default.
- W3048593389 cites W2194775991 @default.
- W3048593389 cites W2217490359 @default.
- W3048593389 cites W2401231614 @default.
- W3048593389 cites W2462831000 @default.
- W3048593389 cites W2518108298 @default.
- W3048593389 cites W2526044239 @default.
- W3048593389 cites W2531409750 @default.
- W3048593389 cites W2546302380 @default.
- W3048593389 cites W2549139847 @default.
- W3048593389 cites W2553303224 @default.
- W3048593389 cites W2612445135 @default.
- W3048593389 cites W2616274621 @default.
- W3048593389 cites W2765407302 @default.
- W3048593389 cites W2766574708 @default.
- W3048593389 cites W2767286248 @default.
- W3048593389 cites W2795247881 @default.
- W3048593389 cites W2796347433 @default.
- W3048593389 cites W2803268044 @default.
- W3048593389 cites W2890166761 @default.
- W3048593389 cites W2896005801 @default.
- W3048593389 cites W2899663614 @default.
- W3048593389 cites W2903382683 @default.
- W3048593389 cites W2914484425 @default.
- W3048593389 cites W2946948417 @default.
- W3048593389 cites W2949117887 @default.
- W3048593389 cites W2953384591 @default.
- W3048593389 cites W2962933129 @default.
- W3048593389 cites W2963125010 @default.
- W3048593389 cites W2963163009 @default.
- W3048593389 cites W2963342610 @default.
- W3048593389 cites W2963420686 @default.
- W3048593389 cites W2963446712 @default.
- W3048593389 cites W2963452022 @default.
- W3048593389 cites W2963587345 @default.
- W3048593389 cites W2963703618 @default.
- W3048593389 cites W2964153283 @default.
- W3048593389 cites W2970206392 @default.
- W3048593389 cites W2970971581 @default.
- W3048593389 cites W2990268359 @default.
- W3048593389 cites W2991015021 @default.
- W3048593389 cites W2992308087 @default.
- W3048593389 cites W2996539502 @default.
- W3048593389 cites W3018757597 @default.
- W3048593389 cites W3118608800 @default.
- W3048593389 hasPublicationYear "2019" @default.
- W3048593389 type Work @default.
- W3048593389 sameAs 3048593389 @default.
- W3048593389 citedByCount "56" @default.
- W3048593389 countsByYear W30485933892019 @default.
- W3048593389 countsByYear W30485933892020 @default.
- W3048593389 countsByYear W30485933892021 @default.
- W3048593389 crossrefType "posted-content" @default.
- W3048593389 hasAuthorship W3048593389A5035144667 @default.
- W3048593389 hasConcept C134306372 @default.
- W3048593389 hasConcept C14036430 @default.
- W3048593389 hasConcept C154945302 @default.
- W3048593389 hasConcept C177264268 @default.
- W3048593389 hasConcept C199360897 @default.
- W3048593389 hasConcept C2776760102 @default.
- W3048593389 hasConcept C2984842247 @default.
- W3048593389 hasConcept C33923547 @default.
- W3048593389 hasConcept C38365724 @default.
- W3048593389 hasConcept C41008148 @default.
- W3048593389 hasConcept C50644808 @default.
- W3048593389 hasConcept C72169020 @default.
- W3048593389 hasConcept C78458016 @default.
- W3048593389 hasConcept C8642999 @default.
- W3048593389 hasConcept C86803240 @default.
- W3048593389 hasConceptScore W3048593389C134306372 @default.
- W3048593389 hasConceptScore W3048593389C14036430 @default.
- W3048593389 hasConceptScore W3048593389C154945302 @default.
- W3048593389 hasConceptScore W3048593389C177264268 @default.
- W3048593389 hasConceptScore W3048593389C199360897 @default.
- W3048593389 hasConceptScore W3048593389C2776760102 @default.
- W3048593389 hasConceptScore W3048593389C2984842247 @default.
- W3048593389 hasConceptScore W3048593389C33923547 @default.
- W3048593389 hasConceptScore W3048593389C38365724 @default.
- W3048593389 hasConceptScore W3048593389C41008148 @default.
- W3048593389 hasConceptScore W3048593389C50644808 @default.