‟Deep lexicography” - fad or opportunity?
In recent years we are witnessing staggering improvements in various intelligent data processing tasks due to the developments in the area of deep learning, ranging from image and video processing to speech processing and natural language understanding. In this talk I want to discuss the opportunities and challenges that these developments pose for the area of electronic lexicography.
The most of my talk I will discuss the concept of representation learning of various elements of language, namely words, lexemes and utterances, and their applicability to lexicography. I will start with the well known approaches to learning static representations of words, the so called word embeddings, and their usage in tasks such as semantic shift detection and prediction of specific lexical features. I will continue with touching upon the multilingual dimension of representation learning and wrap up with the most recent developments in natural language understanding in form of learning dynamic, context-aware representations of words.