Large-scale neural network models combining text and images have made incredible progress in recent years. However, it remains an open question to what extent such models encode compositional representations of the concepts over which they operate, such as correctly identifying ‘red cube’ by reasoning over the constituents ‘red’ and ‘cube’. In this work, we focus on the ability of a large pretrained vision and language model (CLIP) to encode compositional concepts and to bind variables in a structure-sensitive way (e.g., differentiating ‘cube behind sphere’ from ‘sphere behind cube’). To inspect the performance of CLIP, we compare several architectures from research on compositional distributional semantics models (CDSMs), a line of research that attempts to implement traditional compositional linguistic structures within embedding spaces. We benchmark them on three synthetic datasets – single-object, two-object, and relational – designed to test concept binding. We find that CLIP can compose concepts in a single-object setting, but in situations where concept binding is needed, performance drops dramatically. At the same time, CDSMs also perform poorly, with best performance at chance level.

Metaphors in natural language are a reflection of fundamental cognitive processes such as analogical reasoning and categorisation, and are deeply rooted in everyday communication. Metaphor understanding is therefore an essential task for large language models (LLMs). We release the Metaphor Understanding Challenge Dataset (MUNCH), designed to evaluate the metaphor understanding capabilities of LLMs. The dataset provides over 10k paraphrases for sentences containing metaphor use, as well as 1.5k instances containing inapt paraphrases. The inapt paraphrases were carefully selected to serve as control to determine whether the model indeed performs full metaphor interpretation or rather resorts to lexical similarity. All apt and inapt paraphrases were manually annotated. The metaphorical sentences cover natural metaphor uses across 4 genres (academic, news, fiction, and conversation), and they exhibit different levels of novelty. Experiments with LLaMA and GPT-3.5 demonstrate that MUNCH presents a challenging task for LLMs. The dataset is freely accessible at https://github.com/xiaoyuisrain/metaphor-understanding-challenge.

Metaphor is an indispensable part of human cognition and everyday communication. Much research has been conducted elucidating metaphor processing in the mind/brain and the role it plays in communication. in recent years, metaphor processing systems have benefited greatly from these studies, as well as the rapid advances in deep learning for natural language processing (NLP). This paper provides a comprehensive review and discussion of recent developments in automated metaphor processing, in light of the findings about metaphor in the mind, language, and communication, and from the perspective of downstream NLP tasks.

Words can have multiple senses. Compositional distributional models of meaning have been argued to deal well with finer shades of meaning variation known as polysemy, but are not so well equipped to handle word senses that are etymologically unrelated, or homonymy. Moving from vectors to density matrices allows us to encode a probability distribution over different senses of a word, and can also be accommodated within a compositional distributional model of meaning. In this paper we present three new neural models for learning density matrices from a corpus, and test their ability to discriminate between word senses on a range of compositional datasets. When paired with a particular composition method, our best model outperforms existing vector-based compositional models as well as strong sentence encoders.

Metaphors allow us to convey emotion by connecting physical experiences and abstract concepts. The results of previous research in linguistics and psychology suggest that metaphorical phrases tend to be more emotionally evocative than their literal counterparts. In this paper, we investigate the relationship between metaphor and emotion within a computational framework, by proposing the first joint model of these phenomena. We experiment with several multitask learning architectures for this purpose, involving both hard and soft parameter sharing. Our results demonstrate that metaphor identification and emotion prediction mutually benefit from joint learning and our models advance the state of the art in both of these tasks.

Language is used to describe concepts, and many of these concepts are hierarchical. Moreover, this hierarchy should be compatible with forming phrases and sentences. We use linear-algebraic methods that allow us to encode words as collections of vectors. The representations we use have an ordering, related to subspace inclusion, which we interpret as modelling hierarchical information. The word representations built can be understood within a compositional distributional semantic framework, providing methods for composing words to form phrase and sentence level representations. We show that the resulting representations give competitive results on both word-level hyponymy and sentence-level entailment datasets.