Roger Levy

TL;DR: It is argued that researchers using LMEMs for confirmatory hypothesis testing should minimally adhere to the standards that have been in place for many decades, and it is shown thatLMEMs generalize best when they include the maximal random effects structure justified by the design.

TL;DR: A simple information-theoretic characterization of processing difficulty as the work incurred by resource reallocation during parallel, incremental, probabilistic disambiguation in sentence comprehension is proposed, and its equivalence to the theory of Hale is demonstrated.

TL;DR: It is shown that accounting for cross-modal correlations and semantic abstraction both improve retrieval accuracy and are shown to outperform state-of-the-art image retrieval systems on a unimodal retrieval task.

TL;DR: A state-of-the-art computational language model, two large behavioral data-sets, and non-parametric statistical techniques are combined to establish for the first time the quantitative form of the relationship between expectation and reading times, finding that it is logarithmic over six orders of magnitude in estimated predictability.

TL;DR: This work demonstrates that the trend toward predictability-sensitive syntactic reduction (Jaeger, 2006) is robust in the face of a wide variety of control variables, and presents evidence that speakers use both surface and structural cues for predictability estimation.