Discovering Sound Concepts and Acoustic Relations In Text
This work addresses the need for automated acoustic concept extraction from text, which could benefit acoustic event and scene detection research, though it appears incremental in its approach.
The paper tackles the problem of identifying sound-related phrases and defining acoustic scenes from text, using pattern matching, POS tagging, dependency parsing, and LSTM networks to generate sound concepts and predict sets for scenes, with potential applications in acoustic knowledge base creation and event detection.
In this paper we describe approaches for discovering acoustic concepts and relations in text. The first major goal is to be able to identify text phrases which contain a notion of audibility and can be termed as a sound or an acoustic concept. We also propose a method to define an acoustic scene through a set of sound concepts. We use pattern matching and parts of speech tags to generate sound concepts from large scale text corpora. We use dependency parsing and LSTM recurrent neural network to predict a set of sound concepts for a given acoustic scene. These methods are not only helpful in creating an acoustic knowledge base but in the future can also directly help acoustic event and scene detection research.