From Red Wine to Red Tomato: Composition with Context

By Ishan Misra, Abhinav Kumar Gupta, Martial Hebert. 2017

301 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

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

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Ego4D: Around the World in 3,000 Hours of Egocentric Video (2021, background): “[158] Ishan Misra, Abhinav Gupta, and Martial Hebert.”
Cross-Task Weakly Supervised Learning From Instructional Videos (2019, methodology): “It is also related to methods that compose new classifiers from others, including [25, 34, 15] among many others.”
Benchmark for Compositional Text-to-Image Synthesis (2021, background): “Compositionality is a key feature of visual intelligence, and it has been explored in visual perception [28] and image captioning [31] before.”
Small Sample Learning in Big Data Era (2018, methodology): “Through assuming that complex visual concepts could be composed using primitive visual concepts, Misra et al. (2017) presented an approach to compose classifiers to generate classifiers for new complex concepts.”
Adversarial Fine-Grained Composition Learning for Unseen Attribute-Object Recognition (2019, methodology): “We use the compositional split [18], i.”
Task-Driven Modular Networks for Zero-Shot Compositional Learning (2019, methodology): “Most prior approaches to CZSL are based on the idea of embedding the object-attribute pair in image feature space [21, 22].”
Generative Models of Visually Grounded Imagination (2017, background): “Perhaps the most relevant to our paper is (Misra et al., 2017), where they construct a compositional split of their (noun, adjective) data, and learn how to combine individual classifiers for nouns and adjectives to predict novel combinations (this is analogous to our approach”
TAFE-Net: Task-Aware Feature Embeddings for Low Shot Learning (2019, background): “[31].”
Learning to Learn Words from Visual Scenes (2019, background): “tperform. Generalizing to compositions: Due to the diversity of the visual world, there has been extensive work in computer vision on learning compositional representations for objects and attributes [32,22,31,34] as well as for objects and actions [24,34]. In natural language processing, high-level”
Learning Compositional Representations for Few-Shot Recognition (2018, background): “[29] propose to train a network that predicts classifiers for novel concepts by composing existing classifiers for the parts.”
MetaAnchor: Learning to Detect Objects with Customized Anchors (2018, background): “ where weights are predicted by another structure rather than directly learned, which is mainly used in the ﬁelds of learning to learn [9, 1, 41], few/zero-shot learning [4, 41] and transfer learning [26]. For object 2 detection there are a few related works, for example, [14] proposes to predict ma”
Compositional Few-Shot Recognition with Primitive Discovery and Enhancing (2020, background): “o a better interpretability. Compositional recognition is the recognition by primitives, and has been well studied in the cognitive science [6, 15, 26]. This concept has been applied to some domains: [40] points that the model will benefit from compositional learning of visual tasks to have better g”
Deep Mixture of Experts via Shallow Embedding (2018, methodology): “task speciﬁcation. 4.2. Unseen Visual-attribute Composition Besides the standard zero-shot learning benchmarks, we evaluate our model in the visual-attribute composition task proposed by Misra et al. [32]. The goal is to compose a set of visual concept primitives like attributes and objects (e.g. la”
Zero-Shot Recognition via Semantic Embeddings and Knowledge Graphs (2018, methodology): “Given classifiers of primitive visual concepts as inputs, [33] applies a simple composition rule to generate classifiers for new complex concepts.”
Action Modifiers: Learning From Adverbs in Instructional Videos (2019, methodology): “[34] learn how to compose separate object and attributes classifiers for novel combinations.”
Composing Text and Image for Image Retrieval - an Empirical Odyssey (2018, methodology): “To be able to compare to prior work on this dataset, we also consider the classification setting proposed in [31, 33].”