mlsanityData to 7 Oct 2026

Tight Lower Bounds on Worst-Case Guarantees for Zero-Shot Learning with Attributes

arXiv 2205.13068 ↗First posted 25 May 2022cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We develop a rigorous mathematical analysis of zero-shot learning with attributes. In this setting, the goal is to label novel classes with no training data, only detectors for attributes and a description of how those attributes are correlated with the target classes, called the class-attribute matrix. We develop the first non-trivial lower bound on the worst-case error of the best map from attributes to classes for this setting, even with perfect attribute detectors. The lower bound characterizes the theoretical intrinsic difficulty of the zero-shot problem based on the available information -- the class-attribute matrix -- and the bound is practically computable from it. Our lower bound is tight, as we show that we can always find a randomized map from attributes to classes whose expected error is upper bounded by the value of the lower bound. We show that our analysis can be predictive of how standard zero-shot methods behave in practice, including which classes will likely be confused with others.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 27 of its 62 references that are in the map. First posted on arXiv 25 May 2022.

Builds on

27 of 62 references.

References · 27

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  3. 2020Language Models are Few-Shot Learnerscited by 32,199
  4. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,747
  5. 2014Conditional Generative Adversarial Netscited by 2,152
  6. 2021Multitask Prompted Training Enables Zero-Shot Task Generalizationcited by 1,462
  7. 2017Zero-Shot Learning -- A Comprehensive Evaluation of the Good, the Bad and the Uglycited by 536
  8. 2018Taking A Closer Look at Domain Shift: Category-level Adversaries for Semantics Consistent Domain Adaptationcited by 295
  9. 2017Feature Generating Networks for Zero-Shot Learningcited by 257
  10. 2013Zero-Shot Learning Through Cross-Modal Transfercited by 237
  11. 2014Evaluation of Output Embeddings for Fine-Grained Image Classificationcited by 166
  12. 2015Label-Embedding for Image Classificationcited by 164
15 more
  1. 2019f-VAEGAN-D2: A Feature Generating Framework for Any-Shot Learningcited by 164
  2. 2020A Review of Generalized Zero-Shot Learning Methodscited by 159
  3. 2018Zero-shot Recognition via Semantic Embeddings and Knowledge Graphscited by 155
  4. 2016Learning a Deep Embedding Model for Zero-Shot Learningcited by 153
  5. 2017Semantic Autoencoder for Zero-Shot Learningcited by 133
  6. 2018Rethinking Knowledge Graph Propagation for Zero-Shot Learningcited by 87
  7. 2018Multi-modal Cycle-consistent Generalized Zero-Shot Learningcited by 87
  8. 2019Leveraging the Invariant Side of Generative Zero-Shot Learningcited by 83
  9. 2015Zero-Shot Learning via Semantic Similarity Embeddingcited by 74
  10. 2017Unsupervised Domain Adaptation with Similarity Learningcited by 73
  11. 2015Predicting Deep Zero-Shot Convolutional Neural Networks using Textual Descriptionscited by 64
  12. 2014Zero Shot Recognition with Unreliable Attributescited by 39
  13. 2016Predicting Visual Exemplars of Unseen Classes for Zero-Shot Learningcited by 30
  14. 2020Zero-Shot Learning with Common Sense Knowledge Graphscited by 17
  15. 2015Optimally Combining Classifiers Using Unlabeled Datacited by 9

Built on by

Not cited yet.