Ryan Cotterell · page 2
Papers
- A Simple Joint Model for Improved Contextual Neural Lemmatization (2019 · 39 citations)
Lemmatization maps an inflected form to its citation form, and the right lemma often depends on context, as "running" is a verb in one sentence and a noun in another. - Best-First Beam Search (2020 · 28 citations)
Beam search is the standard approximate decoder for neural sequence models such as NMT systems, yet earlier work mostly changed its scoring function rather than its speed. - Translations of the Callhome Egyptian Arabic corpus for conversational speech translation (2014 · 28 citations)
Research on translating conversational speech has been held back by the lack of in-domain data. For Egyptian Arabic in particular, no resource aligned speech, transcripts and English translations from the same domain. As a result, translation systems were usually trained on scripted or written text that differs… - A Cross-Linguistic Pressure for Uniform Information Density in Word Order (2023 · 24 citations)
Languages differ in word order but share statistical tendencies, which are often explained by functional pressures such as dependency length minimization. - On the Proper Treatment of Tokenization in Psycholinguistics (2024 · 23 citations)
Psycholinguists often relate a language model's surprisal for a stretch of text to how hard readers find it, for example via gaze duration. - A Rich Morphological Tagger for English: Exploring the Cross-Linguistic Tradeoff Between Morphology and Syntax (2017 · 21 citations)
The question is whether English, a morphologically poor language, carries in its syntax the same meanings that Czech marks with inflection. - Naturalistic Causal Probing for Morpho-Syntax (2023 · 19 citations)
Conventional (correlational) probes can show that a property is decodable from a model's representations. They cannot show whether the property actually causes those representations, and correlated linguistic features can inflate their scores. The authors propose causal probing through input-level interventions on… - Hexatagging: Projective Dependency Parsing as Tagging (2023 · 16 citations)
The paper turns projective dependency parsing into tagging. Each word receives a "hexatag" from a finite set that encodes the stack actions of a binarized version of its tree, and all words are tagged in parallel with two linear projections over a pretrained encoder. Because validity depends only on the number of… - Can Language Models Learn Typologically Implausible Languages? (2026 · 10 citations)
Languages around the world show strong word-order correlations. For example, verb–object languages usually have prepositions, and object–verb languages usually have postpositions. These patterns are often explained by human learning biases, but it is debated whether such biases are specific to language or arise from… - Towards Developmentally Plausible Rewards: Communicative Success as a Learning Signal for Interactive Language Models (2025 · 4 citations)
Language models learn by predicting the next token, while children also learn from whether their attempts to communicate succeed. - Probing for Reading Times (2026 · 1 citation)
The paper asks whether the hidden states of neural language models hold information about human reading behaviour beyond what the usual single-number predictor, surprisal, provides. - Prefix Parsing is Just Parsing (2026)
Prefix parsing asks whether the beginning of a string can still be completed into a sentence of a context-free grammar.
This page in Markdown → https://md.research.elamur.ai/free/en/authors/ryan-cotterell/page/2/