The Psycho-Biology Of Language

Zipf, George Kingsley

doi:10.4324/9781315009421

Cited by 95 publications

(100 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…where f is the original frequency value and h is the highest frequency found (excluding stop words). This conversion makes use of the Zipfian distribution observed in natural language (Zipf, 1935), allowing the model to be pre-trained on output values that match the range in the shared task dataset (see Section 4.1 for more details). We chose this particular frequency feature because it is the most strongly correlated one with the complexity values in the training data among the 15 features used in the random forest regressor (see Table 1 #12).…”

Section: Frequency Pre-trainingmentioning

confidence: 99%

Cambridge at SemEval-2021 Task 1: An Ensemble of Feature-Based and Neural Models for Lexical Complexity Prediction

Yuan¹,

Tyen²,

Strohmaier³

2021

Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021)

View full text Add to dashboard Cite

This paper describes our submission to the SemEval-2021 shared task on Lexical Complexity Prediction. We approached it as a regression problem and present an ensemble combining four systems, one feature-based and three neural with fine-tuning, frequency pre-training and multi-task learning, achieving Pearson scores of 0.8264 and 0.7556 on the trial and test sets respectively (sub-task 1). We further present our analysis of the results and discuss our findings.

show abstract

Section: Frequency Pre-trainingmentioning

confidence: 99%

Cambridge at SemEval-2021 Task 1: An Ensemble of Feature-Based and Neural Models for Lexical Complexity Prediction

Yuan¹,

Tyen²,

Strohmaier³

2021

Proceedings of the 15th International Workshop on Semantic Evaluation (SemEval-2021)

View full text Add to dashboard Cite

show abstract

“…3 illustrates. In this case, CoHSI predicts that homogeneous systems will be overwhelmingly likely to obey a simple power-law [HW17]; it is in fact a proof of Zipf's law [Zip35]. This too has been validated on multiple qualifying systems to high degrees of confidence [HW17].…”

Section: The Cohsi Modelsmentioning

confidence: 99%

CoHSI IV: Unifying Horizontal and Vertical Gene Transfer - is Mechanism Irrelevant ?

Hatton,

Warr

2018

Preprint

View full text Add to dashboard Cite

In previous papers we have described with strong experimental support, the organising role that CoHSI (Conservation of Hartley-Shannon Information) plays in determining important global properties of all known proteins, from defining the length distribution, to the natural emergence of very long proteins and their relationship to evolutionary time. Here we consider the insight that CoHSI might bring to a different problem, the distribution of identical proteins across species. Horizontal and Vertical Gene Transfer (HGT/VGT) both lead to the replication of protein sequences across species through a diversity of mechanisms some of which remain unknown. In contrast, CoHSI predicts from fundamental theory that such systems will demonstrate power law behavior independently of any mechanisms, and using the Uniprot database we show that the global pattern of protein re-use is emphatically linear on a log-log plot (adj. R 2 = 0.99, p < 2.2 × 10 −16 over 4 decades); i.e. it is extremely close to the predicted power law. Specifically we show that over 6.9 million proteins in TrEMBL 18-02 are re-used, i.e. their sequence appears identically in between 2 and 9,812 species, with re-used proteins varying in length from 7 to as long as 14,596 amino acids. Using (DL+V) to denote the three domains of life plus viruses, 21,676 proteins are shared between two (DL+V); 22 between three (DL+V) and 5 are shared in all four (DL+V). Although the majority of protein re-use occurs between bacterial species those proteins most frequently re-used occur disproportionately in viruses, which play a fundamental role in this distribution.These results suggest that diverse mechanisms of gene transfer (including traditional inheritance) are irrelevant in determining the global distribution of protein re-use.

show abstract

“…The range 1 ≤ r < r min range contains mainly function words. They serve for establishing grammatical constructions (e.g., the, and a, such, this, that, where, were) 3 . The majority of words in the Zipfian range r min < r < r max do have a narrow meaning (content words).…”

Section: Minimal Rank Rminmentioning

confidence: 99%

“…Each such relation acquires the status of law and paves the way towards bringing in methods and ideas of natural sciences. This is why Zipf's law-discovered independently by stenographer Estoup in 1912 [1] and physicist Condon in 1928 [2], and later on advertized by linguist Zipf [3][4][5][6]-attracted so much inter-disciplinary attention. The law applies both to text mixtures (corpora), and to separate texts written in many natural and artificial alphabetic languages [8][9][10], as well as in Chinese characters [11].…”

Section: Introductionmentioning

confidence: 99%

Relating Zipf's law to textual information

Deng¹,

Allahverdyan²

2018

Preprint

View full text Add to dashboard Cite

Zipf's law is the main regularity of quantitative linguistics. Despite of many works devoted to foundations of this law, it is still unclear whether it is only a statistical regularity, or it has deeper relations with information-carrying structures of the text. This question relates to that of distinguishing a meaningful text (written in an unknown system) from a meaningless set of symbols that mimics statistical features of a text. Here we contribute to resolving these questions by comparing features of the first half of a text (from the beginning to the middle) to its second half. This comparison can uncover hidden effects, because the halves have the same values of many parameters (style, genre, author's vocabulary etc). In all studied texts we saw that for the first half Zipf's law applies from smaller ranks than in the second half, i.e. the law applies better to the first half. Also, words that hold Zipf's law in the first half are distributed more homogeneously over the text. These features do allow to distinguish a meaningful text from a random sequence of words. Our findings correlate with a number of textual characteristics that hold in most cases we studied: the first half is lexically richer, has longer and less repetitive words, more and shorter sentences, more punctuation signs and more paragraphs. These differences between the halves indicate on a higher hierarchic level of text organization that so far went unnoticed in text linguistics. They relate the validity of Zipf's law to textual information. A complete description of this effect requires new models, though one existing model can account for some of its aspects.

show abstract

The Psycho-Biology Of Language

Cited by 95 publications

References 0 publications

Cambridge at SemEval-2021 Task 1: An Ensemble of Feature-Based and Neural Models for Lexical Complexity Prediction

Cambridge at SemEval-2021 Task 1: An Ensemble of Feature-Based and Neural Models for Lexical Complexity Prediction

CoHSI IV: Unifying Horizontal and Vertical Gene Transfer - is Mechanism Irrelevant ?

Relating Zipf's law to textual information

Contact Info

Product

Resources

About