Density Matrices for Metaphor Understanding
This work addresses metaphor understanding in natural language processing, but it is incremental as it builds on existing density matrix approaches for lexical ambiguity.
The paper tackled the problem of modeling metaphorical meaning as a type of lexical ambiguity using density matrices, finding that it is more difficult than other ambiguities but that their best method outperformed simple baselines and some neural language models.
In physics, density matrices are used to represent mixed states, i.e. probabilistic mixtures of pure states. This concept has previously been used to model lexical ambiguity. In this paper, we consider metaphor as a type of lexical ambiguity, and examine whether metaphorical meaning can be effectively modelled using mixtures of word senses. We find that modelling metaphor is significantly more difficult than other kinds of lexical ambiguity, but that our best-performing density matrix method outperforms simple baselines as well as some neural language models.