Word embeddings are vectorial semantic representations built with either counting or predicting techniques aimed at capturing shades of meaning from word co-occurrences. Since their introduction, these representations have been criticized for lacking interpretable dimensions. This property of word embeddings limits our understanding of the semantic features they actually encode. Moreover, it contributes to the “black box” nature of the tasks in which they are used, since the reasons for word embedding performance often remain opaque to humans. In this contribution, we explore the semantic properties encoded in word embeddings by mapping them onto interpretable vectors, consisting of explicit and neurobiologically motivated semantic features (Binder et al. 2016). Our exploration takes into account different types of embeddings, including factorized count vectors and predict models (Skip-Gram, GloVe, etc.), as well as the most recent contextualized representations (i.e., ELMo and BERT).In our analysis, we first evaluate the quality of the mapping in a retrieval task, then we shed light on the semantic features that are better encoded in each embedding type. A large number of probing tasks is finally set to assess how the original and the mapped embeddings perform in discriminating semantic categories. For each probing task, we identify the most relevant semantic features and we show that there is a correlation between the embedding performance and how they encode those features. This study sets itself as a step forward in understanding which aspects of meaning are captured by vector spaces, by proposing a new and simple method to carve human-interpretable semantic representations from distributional vectors.

Modality exclusivity norms have been developed in different languages for research on the relationship between perceptual and conceptual systems. This paper sets up the first modality exclusivity norms for Chinese, a Sino-Tibetan language with semantics as its orthographically relevant level. The norms are collected through two studies based on Chinese sensory words. The experimental designs take into consideration the morpho-lexical and orthographic structures of Chinese. Study 1 provides a set of norms for Mandarin Chinese single-morpheme words in mean ratings of the extent to which a word is experienced through the five sense modalities. The degrees of modality exclusivity are also provided. The collected norms are further analyzed to examine how sub-lexical orthographic representations of sense modalities in Chinese characters affect speakers’ interpretation of the sensory words. In particular, we found higher modality exclusivity rating for the sense modality explicitly represented by a semantic radical component, as well as higher auditory dominant modality rating for characters with transparent phonetic symbol components. Study 2 presents the mean ratings and modality exclusivity of coordinate disyllabic compounds involving multiple sense modalities. These studies open new perspectives in the study of modality exclusivity. First, links between modality exclusivity and writing systems have been established which has strengthened previous accounts of the influence of orthography in the processing of visual information in reading. Second, a new set of modality exclusivity norms of compounds is proposed to show the competition of influence on modality exclusivity from different linguistic factors and potentially allow such norms to be linked to studies on synesthesia and semantic transparency.

Attention models are proposed in sentiment analysis and other classification tasks because some words are more important than others to train the attention models. However, most existing methods either use local context based information, affective lexicons, or user preference information. In this work, we propose a novel attention model trained by cognition grounded eye-tracking data. First,a reading prediction model is built using eye-tracking data as dependent data and other features in the context as independent data. The predicted reading time is then used to build a cognition grounded attention layer for neural sentiment analysis. Our model can capture attentions in context both in terms of words at sentence level as well as sentences at document level. Other attention mechanisms can also be incorporated together to capture other aspects of attentions, such as local attention, and affective lexicons. Results of our work include two parts. The first part compares our proposed cognition ground attention model with other state-of-the-art sentiment analysis models. The second part compares our model with an attention model based on other lexicon based sentiment resources. Evaluations show that sentiment analysis using cognition grounded attention model outperforms the state-of-the-art sentiment analysis methods significantly. Comparisons to affective lexicons also indicate that using cognition grounded eye-tracking data has advantages over other sentiment resources by considering both word information and context information. This work brings insight to how cognition grounded data can be integrated into natural language processing (NLP) tasks.