paper-with-me

Papers

Arithmetics-Based Decomposition of Numeral Words -- Arithmetic Conditions give the Unpacking Strategy

2023-12-14 · Isidor Konrad Maier, Matthias Wolff

This paper presents a novel numeral decomposer based on arithmetic criteria. The criteria are not dependent on a base-10 assumption but only on Hurford's Packing Strategy. Hurford's Packing Strategy constitutes numerals by packing factors and summands to multiplicators. We found out that a numeral of value n has a multiplicator larger than sqrt(n), a summand smaller than n/2 and a factor smaller than sqrt(n). Using these findings, the numeral decomposer attempts to detect and unpack factors and summand in order to reverse Hurford's Packing strategy. We tested its applicability for incremental unsupervised grammar induction in 273 languages. This way, grammars were obtained with sensible mathematical attributes that explain the structure of produced numerals. The numeral-decomposer-induced grammars are often close to expert-made and more compact than numeral grammars induced by a modern state-of-the-art grammar induction tool. Furthermore, this paper contains a report about the few cases of incorrect induced mathematical attributes, which are often linked to linguistic peculiarities like context sensitivity.

📄 PDF Abstract BibTeX arXiv:2312.10097

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

BASE 설명 없음

Similar Papers 제목 키워드 기반

Multiplication in Multimodal LLMs: Computation with Text, Image, and Audio Inputs

2026-04-20 · Samuel G. Balter, Ethan Jerzak, Connor T. Jerzak arxiv

Multimodal LLMs can accurately perceive numerical content across modalities yet fail to perform exact multi-digit multiplication when the identical underlying arithmetic problem is presented as numerals, number words, im…

tntorch: Tensor Network Learning with PyTorch

2022-06-22 · Mikhail Usvyatsov, Rafael Ballester-Ripoll, Konrad Schindler

We present tntorch, a tensor learning framework that supports multiple decompositions (including Candecomp/Parafac, Tucker, and Tensor Train) under a unified interface. With our library, the user can learn and handle low…

GPUtensor algebra

Deep Neural Network inference with reduced word length

2018-10-23 · Lukas Mauch, Bin Yang

Deep neural networks (DNN) are powerful models for many pattern recognition tasks, yet their high computational complexity and memory requirement limit them to applications on high-performance computing platforms. In thi…

Spectral Graph-Based Method of Multimodal Word Embedding

2017-08-01 · WS 2017 8 · Kazuki Fukui, Takamasa Oshikiri, Hidetoshi Shimodaira

In this paper, we propose a novel method for multimodal word embedding, which exploit a generalized framework of multi-view spectral graph embedding to take into account visual appearances or scenes denoted by words in a…

Graph EmbeddingImage RetrievalMachine TranslationPart-Of-Speech Tagging+4

Reinforcement Learning of Minimalist Numeral Grammars

2019-06-11 · Peter beim Graben, Ronald Römer, Werner Meyer, Markus Huber 외

Speech-controlled user interfaces facilitate the operation of devices and household functions to laymen. State-of-the-art language technology scans the acoustically analyzed speech signal for relevant keywords that are s…

reinforcement-learningReinforcement LearningReinforcement Learning (RL)slot-filling+1