Corpus ID: 199501735

Do Neural Language Representations Learn Physical Commonsense?

@article{Forbes2019DoNL,
  title={Do Neural Language Representations Learn Physical Commonsense?},
  author={M. Forbes and Ari Holtzman and Yejin Choi},
  journal={ArXiv},
  year={2019},
  volume={abs/1908.02899}
}
  • M. Forbes, Ari Holtzman, Yejin Choi
  • Published 2019
  • Computer Science, Psychology
  • ArXiv
  • Humans understand language based on the rich background knowledge about how the physical world works, which in turn allows us to reason about the physical world through language. [...] Key Result While recent advancements of neural language models have demonstrated strong performance on various types of natural language inference tasks, our study based on a dataset of over 200k newly collected annotations suggests that neural language representations still only learn associations that are explicitly written…Expand Abstract

    Figures, Tables, and Topics from this paper.

    Explore Further: Topics Discussed in This Paper

    Can a Gorilla Ride a Camel? Learning Semantic Plausibility from Text
    • 1
    • Open Access
    Inducing Relational Knowledge from BERT
    • 15
    • Open Access
    ESPRIT: Explaining Solutions to Physical Reasoning Tasks
    • 2
    • Open Access
    Ecological Semantics: Programming Environments for Situated Language Understanding
    • 1
    • Open Access
    Probing Text Models for Common Ground with Visual Representations
    • 3
    • Open Access
    How Pre-trained Word Representations Capture Commonsense Physical Comparisons
    • 1
    • Open Access
    Language (Re)modelling: Towards Embodied Language Understanding
    • 1
    • Open Access

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 23 REFERENCES
    Human few-shot learning of compositional instructions
    • 27
    • Open Access
    BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
    • 9,745
    • Open Access
    Deep contextualized word representations
    • 4,082
    • Highly Influential
    • Open Access
    Extracting implicit knowledge from text
    • 15
    Glove: Global Vectors for Word Representation
    • 14,464
    • Open Access
    Semantic feature production norms for a large set of living and nonliving things
    • 791
    • Highly Influential
    • Open Access
    Adam: A Method for Stochastic Optimization
    • 49,106
    • Open Access