Publications

Abstract

Tokenization significantly influences language models (LMs)’ performance. This paper traces the evolution of tokenizers from word-level to subword-level, analyzing how they balance tokens and types to enhance model adaptability while controlling complexity. Despite subword tokenizers like Byte Pair Encoding (BPE) overcoming many word tokenizer limitations, they encounter difficulties in handling non-Latin languages and depend heavily on extensive training data and computational resources to grasp the nuances of multiword expressions (MWEs). This article argues that tokenizers, more than mere technical tools, should drawing inspiration from the cognitive science about human language processing. This study then introduces the “Principle of Least Effort” from cognitive science, that humans naturally seek to reduce cognitive effort, and discusses the benefits of this principle for tokenizer development. Based on this principle, the paper proposes that the Less-is-Better (LiB) model could be a new approach for LLM tokenizer. The LiB model can autonomously learn an integrated vocabulary consisting of subwords, words, and MWEs, which effectively reduces both the numbers of tokens and types. Comparative evaluations show that the LiB tokenizer outperforms existing word and BPE tokenizers, presenting an innovative method for tokenizer development, and hinting at the possibility of future cognitive science-based tokenizers being more efficient.

Abstract

Words typically form the basis of psycholinguistic and computational linguistic studies about sentence processing. However, recent evidence shows the basic units during reading, i.e., the items in the mental lexicon, are not always words, but could also be sub-word and supra-word units. To recognize these units, human readers require a cognitive mechanism to learn and detect them. In this paper, we assume eye fixations during reading reveal the locations of the cognitive units, and that the cognitive units are analogous with the text units discovered by unsupervised segmentation models. We predict eye fixations by model-segmented units on both English and Dutch text. The results show the model-segmented units predict eye fixations better than word units. This finding suggests that the predictive performance of model-segmented units indicates their plausibility as cognitive units. The Less-is-Better (LiB) model, which finds the units that minimize both long-term and working memory load, offers advantages both in terms of prediction score and efficiency among alternative models. Our results also suggest that modeling the least-effort principle for the management of long-term and working memory can lead to inferring cognitive units. Overall, the study supports the theory that the mental lexicon stores not only words but also smaller and larger units, suggests that fixation locations during reading depend on these units, and shows that unsupervised segmentation models can discover these units.

Abstract

Results from a recent neuroimaging study on spoken sentence comprehension have been interpreted as evidence for cortical entrainment to hierarchical syntactic structure. We present a simple computational model that predicts the power spectra from this study, even
though the model's linguistic knowledge is restricted to the lexical level, and word-level representations are not combined into higher-level units (phrases or sentences). Hence, the
cortical entrainment results can also be explained from the lexical properties of the stimuli, without recourse to hierarchical syntax.

Abstract

Electroencephalography (EEG) provides high temporal resolution cognitive information from non-invasive recordings. However, one of the common practices-using a subset of sensors in ERP analysis is hard to provide a holistic and precise dynamic results. Selecting or grouping subsets of sensors may also be subject to selection bias, multiple comparison, and further complicated by individual differences in the group-level analysis. More importantly, changes in neural generators and variations in response magnitude from the same neural sources are difficult to separate, which limit the capacity of testing different aspects of cognitive hypotheses. We introduce EasyEEG, a toolbox that includes several multivariate analysis methods to directly test cognitive hypotheses based on topographic responses that include data from all sensors. These multivariate methods can investigate effects in the dimensions of response magnitude and topographic patterns separately using data in the sensor space, therefore enable assessing neural response dynamics. The concise workflow and the modular design provide user-friendly and programmer-friendly features. Users of all levels can benefit from the open-sourced, free EasyEEG to obtain a straightforward solution for efficient processing of EEG data and a complete pipeline from raw data to final results for publication.