Living a discrete life in a continuous world: Reference with distributed representations
This work addresses a fundamental challenge in natural language processing for AI systems, though it is incremental as it builds on existing external memory models.
The paper tackles the problem of modeling reference in language, which involves handling both continuous and discrete aspects of meaning, by introducing a cross-modal entity tracking task and proposing a neural network architecture with external memory. The model beats traditional neural network architectures on the task but is outperformed by Memory Networks.
Reference is a crucial property of language that allows us to connect linguistic expressions to the world. Modeling it requires handling both continuous and discrete aspects of meaning. Data-driven models excel at the former, but struggle with the latter, and the reverse is true for symbolic models. This paper (a) introduces a concrete referential task to test both aspects, called cross-modal entity tracking; (b) proposes a neural network architecture that uses external memory to build an entity library inspired in the DRSs of DRT, with a mechanism to dynamically introduce new referents or add information to referents that are already in the library. Our model shows promise: it beats traditional neural network architectures on the task. However, it is still outperformed by Memory Networks, another model with external memory.