Representing Numbers in NLP: a Survey and a Vision

By Avijit Thawani, Jay Pujara, Pedro A. Szekely, Filip Ilievski. 2021

147 citations recorded by the source corpus.

NLP systems rarely give special consideration to numbers found in text. This starkly contrasts with the consensus in neuroscience that, in the brain, numbers are represented differently from words. We arrange recent NLP work on numeracy into a comprehensive taxonomy of tasks and methods. We break down the subjective notion of numeracy into 7 subtasks, arranged along two dimensions: granularity (exact vs approximate) and units (abstract vs grounded). We analyze the myriad representational choices made by over a dozen previously published number encoders and decoders. We synthesize best practices for representing numbers in text and articulate a vision for holistic numeracy in NLP, comprised of design trade-offs and a unified evaluation.

Subfields: Artificial Intelligence, Natural Language Processing

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 Natural Language Processing

Citations in the corpus

Survey of Hallucination in Natural Language Generation (2022, background): “However, the correctness of numerals in generated text, such as date, quantities and scalars are important for readers [180, 230, 234].”
Investigating the Limitations of Transformers with Simple Arithmetic Tasks (2021, methodology): “A separate development set of 1,000 examples was used to select the best checkpoint of each run.”
Teaching Algorithmic Reasoning via In-context Learning (2022, methodology): “…the architecture align to the algorithms under consideration (Kaiser and Sutskever, 2015; Chiang and Chen, 2018; Xu et al., 2019; Gordon et al., 2019; Yan et al., 2020; Chen et al., 2020; Xhonneux et al., 2021; Veliˇckovi´c and Blundell, 2021) or to the input format (Thawani et al., 2021).”
Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedback (2023, background): “Language models can express their uncertainty with numerical probabilities as well or better than with words , which is surprising in light of long-standing difficulties in representing numbers in language models (Thawani et al., 2021).”
Evaluation and mitigation of the limitations of large language models in clinical decision-making (2024, unknown): “The difficulty LLMs have in interpreting numbers 48 and solving simple arithmetic 49 is an active research topic 50,51 .”
A Survey of Deep Learning for Mathematical Reasoning (2022, background): “, 2019; Ughade and Kumbhar, 2019), representing numbers representation (Thawani et al., 2021), or”
Embedding Models Measure in Peculiar Ways (2026, unknown): “To conclude, we wish to emphasize what Thawani et al. (2021)”