Large-Scale Object Classification Using Label Relation Graphs

By Jia Deng, Nan Ding, Yangqing Jia, Andrea Frome, Kevin P. Murphy, Samy Bengio, Yuan Li, Hartmut Neven, Hartwig Adam. 2014

481 citations recorded by the source corpus.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Optimization Methods for Large-Scale Machine Learning (2016, neutral): “, see [62, 46].”
Examining CNN representations with respect to Dataset Bias (2017, methodology): “Given a feature map produced by a CNN, Dosovitskiy et al. (Dosovitskiy and Brox 2016) trained a new up-convolutional network to invert the feature map to the original image.”
Computational Baby Learning (2014, background): “Recent successes in computer vision [23] [14] [27] [4], however, largely rely on the big number of labeled instances of visual concepts, which may require considerable human efforts.”
Discovering states and transformations in image collections (2015, background): “Similarly to previous works on transfer learning [6, 4, 19, 28] , our underlying assumption is the transferrability of knowledge between adjectives (state and transformation) (see Fig 2).”
Multi-Task CNN Model for Attribute Prediction (2015, background): “On the other hand, most of the related works [21], [45] tend to apply multi-task learning to allow sharing or using some label relationship heuristics a priori [11].”
Semi-supervised Vocabulary-Informed Learning (2016, unknown): “5 HEX [9] A CNNDECAF 44.”
Multiple Granularity Descriptors for Fine-Grained Categorization (2015, background): “From that perspective, previous work exploring similar idea is [9], which introduces Hierarchy and Exclusion (HEX) graphs to capture semantic relations between labels instead.”
Iterative Visual Reasoning Beyond Convolutions (2018, background): “As an active research area, there are recent works also applying neural networks to the graph structured data [42, 17, 27, 23, 35, 7, 32], or attempting to regularize the output of networks with relationships [8] and knowledge bases [20].”
Large-Scale Visual Relationship Understanding (2018, background): “The second challenge lies in the fact that object categories are often semantically associated[5, 16, 4], and such connections could be more subtle for relationships.”
Beyond One-hot Encoding: lower dimensional target embedding (2018, methodology): “In the field of large-scale recognition, hierarchical approaches such as using tree-based priors [38], label relational graphs [11], CNN hierarchies [46], and HD-CNNs [47] have been proposed.”
Multi-label Zero-Shot Learning with Structured Knowledge Graphs (2017, background): “The Hierarchy and Exclusion (HEX) graph [10] captures semantic relations: mutual exclusion, overlap and subsumption between any two labels, improving object classification by exploiting the label relations.”
Small Sample Learning in Big Data Era (2018, methodology): “Specifically, the semantic relationship between the seen and unseen classes is modelled by the distance between their prototypes to combine knowledge of seen classs, or the knowledge graph to distill the relationships (Rohrbach et al., 2010; Fergus et al., 2010; Deng et al., 2014; Marino et al., 201”
Unsupervised Domain Adaptation for Zero-Shot Learning (2015, background): “Extensive experiments on four challenging object and action benchmark datasets [18, 34, 32] validate the advantages of the proposed model, and demonstrate that the proposed model outperforms the state-of-the-arts naive transfer learning based models [4, 18, 1, 10, 37, 15, 19, 20, 11, 2] when applied”
Knowledge-Embedded Routing Network for Scene Graph Generation (2019, background): “Deng et al. [7] introduced semantic relations including mutual exclusion, overlap, and subsumption, as constraints in the loss function to train the classiﬁers.”
Scene Graph Generation With External Knowledge and Image Reconstruction (2019, background): “There has been growing interest in improving data-driven models with external Knowledge Bases (KBs) in natural language processing [17, 4] and computer vision communities [24, 1, 6].”