Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence WSABIE: Scaling Up To Large Vocabulary Image Annotation

Subfields: Computer Vision and Pattern Recognition, Information Extraction

Citations in the corpus

• Distributed Representations of Words and Phrases and their Compositionality (2013, background): “The follow up work includes applications to automatic speech recognition and machine translation [14, 7], and a wide range of NLP tasks [2, 20, 15, 3, 18, 19, 9].”
• A Neural Network for Factoid Question Answering over Paragraphs (2014, methodology): “Since rank(c, s, Z) is expensive to compute, we approximate it by randomly sampling K incorrect answers until a violation is observed (xc · hs < 1 + xz · hs) and set rank(c, s, Z) = (|Z|−1)/K, as in previous work (Weston et al., 2011; Hermann et al., 2014).”
• Memory Networks (2014, background): “This then gives a feature space of D = 8|W |.”
• Self-Supervised MultiModal Versatile Networks (2020, background): “WSABIE [86] and DeVise [24] introduced the idea of embedding text and image in the same space.”
• Learning Visual Features from Large Weakly Supervised Data (2015, methodology): “In preliminary experiments, we also considered a pairwise ranking loss [50, 53].”
• Predicting Deep Zero-Shot Convolutional Neural Networks Using Textual Descriptions (2015, background): “In [29], this is accomplished via a ranking objective using linear projections of image features and bag-of-words attribute features.”
• Elucidating image-to-set prediction: An analysis of models, losses and datasets (2019, background): “Other solutions to capture label co-occurrences include learning joint input and label embeddings with ranking-based losses [53, 26, 56, 24, 12] as well as designing loss functions such as target distribution mean squared error [50] or target distribution cross-entropy [12, 32, 41], which directly a”
• Bag of Tricks for Efficient Text Classification (2016, methodology): “We also compare with Tagspace (Weston et al., 2014), which is a tag prediction model similar to ours, but based on the Wsabie model of Weston et al. (2011).”
• Building high-level features using large scale unsupervised learning (2011, background): “This result is interesting, but unfortunately requires a certain degree of supervision during dataset construction: their training images (i.e., Caltech 101 images) are aligned, homogeneous and belong to one selected category.”
• Learning Deep Structure-Preserving Image-Text Embeddings (2015, background): “This results in a model similar to WSABIE [49].”
• Multimodal Compact Bilinear Pooling for Visual Question Answering and Visual Grounding (2016, background): “…et al., 2004) e.g. (Gong et al., 2014; Klein et al., 2015; Plummer et al., 2015), linear models with ranking loss (Frome et al., 2013; Karpathy and Fei-Fei, 2015; Socher et al., 2014; Weston et al., 2011) or non-linear deep learning models (Kiros et al., 2014; Mao et al., 2015; Ngiam et al., 2011).”
• Fusion of Detected Objects in Text for Visual Question Answering (2019, neutral): “A notable example of a dual encoder for image classification is WSABIE, proposed by Weston et al. (2011).”
• Deep Unordered Composition Rivals Syntactic Methods for Text Classification (2015, methodology): “For fair comparison with QANTA , we use a normalized tanh activation function at the last layer instead of ReLu, and we also change the output layer from a softmax to the margin ranking loss (Weston et al., 2011) used in QANTA .”
• A Multi-View Embedding Space for Modeling Internet Images, Tags, and Their Semantics (2012, methodology): “Weston et al. (2011) evaluate their Wsabie annotation system on millions of images.”
• Variational Autoencoders for Collaborative Filtering (2018, background): “Top-N ranking loss is difficult to optimize directly and previous work on direct ranking loss minimization resorts to relaxations and approximations [49, 50].”
• Deep Neural Networks for YouTube Recommendations (2016, methodology): “Our deep collaborative filtering model is able to effectively assimilate many signals and model their interaction with layers of depth, outperforming previous matrix factorization approaches used at YouTube [23].”
• Gaussian Ranking by Matrix Factorization (2015, background): “Moreover, weighting schemes [19, 20] and sampling schemes for the negative items [21] were proposed, and may be viewed as extensions to BPR. Incontrast, the presented approach considers all items concerning a user u at once, rather than each user-item pair ( u, i ) separately.”