Matches in SemOpenAlex for { <https://semopenalex.org/work/W3211930213> ?p ?o ?g. }
- W3211930213 abstract "A visual hard attention model actively selects and observes a sequence of subregions in an image to make a prediction. The majority of hard attention models determine the attention-worthy regions by first analyzing a complete image. However, it may be the case that the entire image is not available initially but instead sensed gradually through a series of partial observations. In this paper, we design an efficient hard attention model for classifying such sequentially observed scenes. The presented model never observes an image completely. To select informative regions under partial observability, the model uses Bayesian Optimal Experiment Design. First, it synthesizes the features of the unobserved regions based on the already observed regions. Then, it uses the predicted features to estimate the expected information gain (EIG) attained, should various regions be attended. Finally, the model attends to the actual content on the location where the EIG mentioned above is maximum. The model uses a) a recurrent feature aggregator to maintain a recurrent state, b) a linear classifier to predict the class label, c) a Partial variational autoencoder to predict the features of unobserved regions. We use normalizing flows in Partial VAE to handle multi-modality in the feature-synthesis problem. We train our model using a differentiable objective and test it on five datasets. Our model gains 2-10% higher accuracy than the baseline models when both have seen only a couple of glimpses." @default.
- W3211930213 created "2021-11-22" @default.
- W3211930213 creator A5065422451 @default.
- W3211930213 creator A5079264223 @default.
- W3211930213 date "2021-11-15" @default.
- W3211930213 modified "2023-09-27" @default.
- W3211930213 title "A Probabilistic Hard Attention Model For Sequentially Observed Scenes." @default.
- W3211930213 cites W1484210532 @default.
- W3211930213 cites W1497599070 @default.
- W3211930213 cites W1514535095 @default.
- W3211930213 cites W1522301498 @default.
- W3211930213 cites W1724369340 @default.
- W3211930213 cites W1846479601 @default.
- W3211930213 cites W1956526898 @default.
- W3211930213 cites W1975998725 @default.
- W3211930213 cites W2014018052 @default.
- W3211930213 cites W2014114527 @default.
- W3211930213 cites W2054802006 @default.
- W3211930213 cites W2060442802 @default.
- W3211930213 cites W2076580309 @default.
- W3211930213 cites W2082665022 @default.
- W3211930213 cites W2092143568 @default.
- W3211930213 cites W2098737324 @default.
- W3211930213 cites W2102179764 @default.
- W3211930213 cites W2115441154 @default.
- W3211930213 cites W2115595010 @default.
- W3211930213 cites W2115812046 @default.
- W3211930213 cites W2120078010 @default.
- W3211930213 cites W2120889539 @default.
- W3211930213 cites W2128272608 @default.
- W3211930213 cites W2135440260 @default.
- W3211930213 cites W2138068405 @default.
- W3211930213 cites W2141399712 @default.
- W3211930213 cites W2147527908 @default.
- W3211930213 cites W2154071538 @default.
- W3211930213 cites W2187089797 @default.
- W3211930213 cites W2188365844 @default.
- W3211930213 cites W2335728318 @default.
- W3211930213 cites W2409550820 @default.
- W3211930213 cites W2587284713 @default.
- W3211930213 cites W2604195031 @default.
- W3211930213 cites W2613718673 @default.
- W3211930213 cites W2738588019 @default.
- W3211930213 cites W2894458036 @default.
- W3211930213 cites W2895171208 @default.
- W3211930213 cites W2903048756 @default.
- W3211930213 cites W2943987826 @default.
- W3211930213 cites W2948627722 @default.
- W3211930213 cites W2949117887 @default.
- W3211930213 cites W2949417160 @default.
- W3211930213 cites W2951004968 @default.
- W3211930213 cites W2962741254 @default.
- W3211930213 cites W2962853854 @default.
- W3211930213 cites W2962908092 @default.
- W3211930213 cites W2963139417 @default.
- W3211930213 cites W2963255313 @default.
- W3211930213 cites W2963386218 @default.
- W3211930213 cites W2963420272 @default.
- W3211930213 cites W2963951231 @default.
- W3211930213 cites W2964069537 @default.
- W3211930213 cites W2970898247 @default.
- W3211930213 cites W2990859148 @default.
- W3211930213 cites W3008062244 @default.
- W3211930213 cites W3011955763 @default.
- W3211930213 cites W3025348309 @default.
- W3211930213 cites W3034482833 @default.
- W3211930213 cites W3035296770 @default.
- W3211930213 cites W3043547428 @default.
- W3211930213 cites W3118608800 @default.
- W3211930213 cites W3132534904 @default.
- W3211930213 cites W3146838169 @default.
- W3211930213 hasPublicationYear "2021" @default.
- W3211930213 type Work @default.
- W3211930213 sameAs 3211930213 @default.
- W3211930213 citedByCount "0" @default.
- W3211930213 crossrefType "posted-content" @default.
- W3211930213 hasAuthorship W3211930213A5065422451 @default.
- W3211930213 hasAuthorship W3211930213A5079264223 @default.
- W3211930213 hasConcept C101738243 @default.
- W3211930213 hasConcept C108583219 @default.
- W3211930213 hasConcept C115961682 @default.
- W3211930213 hasConcept C119857082 @default.
- W3211930213 hasConcept C134306372 @default.
- W3211930213 hasConcept C138885662 @default.
- W3211930213 hasConcept C153180895 @default.
- W3211930213 hasConcept C154945302 @default.
- W3211930213 hasConcept C202615002 @default.
- W3211930213 hasConcept C2776401178 @default.
- W3211930213 hasConcept C28826006 @default.
- W3211930213 hasConcept C33923547 @default.
- W3211930213 hasConcept C36299963 @default.
- W3211930213 hasConcept C41008148 @default.
- W3211930213 hasConcept C41895202 @default.
- W3211930213 hasConcept C49937458 @default.
- W3211930213 hasConcept C95623464 @default.
- W3211930213 hasConceptScore W3211930213C101738243 @default.
- W3211930213 hasConceptScore W3211930213C108583219 @default.
- W3211930213 hasConceptScore W3211930213C115961682 @default.
- W3211930213 hasConceptScore W3211930213C119857082 @default.
- W3211930213 hasConceptScore W3211930213C134306372 @default.