On Evaluation Metrics in Optimality Theory

Rasin, Ezer; Katzir, Roni

doi:10.1162/ling_a_00210

Cited by 31 publications

(33 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The argument for bare MDL as the null hypothesis can be taken to support approaches in the literature that use MDL for learning, such as the works mentioned in Section 2.2.3, and in particular works such as de Marcken (1996) and Rasin and Katzir (2013) that use MDL not simply as a convenient heuristic but as the sole principle that maps an explicit UG to an evaluation metric.…”

Section: 19mentioning

confidence: 88%

“…MDL has been used for grammar induction in the works of Berwick (1982), Rissanen and Ristad (1994), Stolcke (1994), Brent and Cartwright (1996), Chen (1996), Grünwald (1996), de Marcken (1996, Osborne and Briscoe (1997), Brent (1999), Clark (2001), Goldsmith (2001), Onnis et al (2002), Zuidema (2003), Dowman (2007), Chang (2008), and Rasin and Katzir (2013) among others. In Section 3.3 I will suggest that MDL arises as a natural criterion for the evaluation of grammars given the data -and thus as a natural CG learning mechanism -from the commitment to an explicit UG made in TL.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A cognitively plausible model for grammar induction

Katzir

2015

JLM

Self Cite

View full text Add to dashboard Cite

This paper aims to bring theoretical linguistics and cognition-general theories of learning into closer contact. I argue that linguists' notions of rich UGs are well-founded, but that cognition-general learning approaches are viable as well and that the two can and should co-exist and support each other. Specifically, I use the observation that any theory of UG provides a learning criterion -- the total memory space used to store a grammar and its encoding of the input -- that supports learning according to the principle of Minimum Description-Length. This mapping from UGs to learners maintains a minimal ontological commitment: the learner for a particular UG uses only what is already required to account for linguistic competence in adults. I suggest that such learners should be our null hypothesis regarding the child's learning mechanism, and that furthermore, the mapping from theories of UG to learners provides a framework for comparing theories of UG.

show abstract

Section: 19mentioning

confidence: 88%

mentioning

confidence: 99%

A cognitively plausible model for grammar induction

Katzir

2015

JLM

Self Cite

View full text Add to dashboard Cite

show abstract

“…In generative linguistics, the idea that simplicity could be used as a learning criterion for the child was explored in Halle (1962), Chomsky (1965), and Chomsky and Halle (1968) for both phonology and syntax. Simplicity-based learning criteria have later been proposed for a variety of language-learning tasks by Berwick (1982), de Marcken (1996, Goldsmith (2001), Chater and Vitányi (2007), Rasin and Katzir (2016), and many others.…”

Section: Discussionmentioning

confidence: 99%

The nature of the semantic stimulus: the acquisition of every as a case study

Rasin

Aravind

2020

Nat Lang Semantics

Self Cite

View full text Add to dashboard Cite

We evaluate the richness of the child’s input in semantics and its relation to the hypothesis space available to the child. Our case study is the acquisition of the universal quantifier every. We report two main findings regarding the acquisition of every on the basis of a corpus study of child-directed and child-ambient speech. Our first finding is that the input in semantics (as opposed to the input in syntax or phonology) is rich enough to systematically eliminate instances of the subset problem of language acquisition: overly general hypotheses about the meaning of every can violate pragmatic constraints, making such hypotheses incompatible with the child’s input. Our second finding is that the semantic input is too poor to eliminate instances of what we refer to as the superset problem, the mirror image of the subset problem. We argue that at least some overly specific hypotheses about the meaning of every are compatible with the child’s input, suggesting either that those hypotheses are not made available by UG or that non-trivial inductive biases are involved in children’s acquisition of every.

show abstract

“…Our first reason for finding MDL promising is that it has provided working distributional learners for phonology -that is, implemented algorithms that take raw surface data and induce a phonology and a lexicon. The MDL learner of Rasin & Katzir (2016), for example, takes unanalyzed surface forms and induces a lexicon, often with abstract URs that differ from the surface forms, along with a set of markedness and faithfulness constraints and their ranking. 7 The MDL learner of Rasin et al (2018a) and Rasin et al (2018b) accomplishes a similar task but within rule-based phonology: it takes unanalyzed surface forms and induces a lexicon and a set of ordered context-sensitive rewrite rules.…”

Section: Motivation For Testing the Predictions Of MDLmentioning

confidence: 99%

A conditional learnability argument for constraints on underlying representations

Rasin

Katzir

2020

J. Ling.

Self Cite

View full text Add to dashboard Cite

We explore the implications of a particular approach to learning for an architectural question in phonology. The learning approach follows the principle of Minimum Description Length (MDL), which has recently been used for learning in both constraint-based and rule-based phonology. The architectural question on which we focus is whether the grammar allows language-specific statements to be made at the level of the lexicon, as was assumed in early generative phonology, or whether such statements are prohibited, as is commonly assumed within more recent work. We show that under MDL, the architectural question has real empirical implications: across a range of seemingly natural representational schemes, an ability to make language-specific statements about the lexicon is needed to ensure the learnability of an important aspect of phonological knowledge.

show abstract

On Evaluation Metrics in Optimality Theory

Cited by 31 publications

References 51 publications

A cognitively plausible model for grammar induction

A cognitively plausible model for grammar induction

The nature of the semantic stimulus: the acquisition of every as a case study

A conditional learnability argument for constraints on underlying representations

Contact Info

Product

Resources

About