Cognitive nodes among lines of lyrics | Michael Kai Petersen, Lars Kai Hansen
| | Abstract | The challenge of recommending songs when confronted with the vast number of music tracks and video streams available on the internet, might require new types of cognitive interfaces reflecting how we emotionally perceive media content. Both in music and language we rely on syntax for parsing sequences of symbols, which based on hierarchically nested structures allow us to express and share the meaning contained within a sentence or a melodic phrase. These structures become part of our memories when the `bottom-up' sensory input raises above the background noise of core affect, and `top-down' trigger distinct emotions reflecting a shift of our attention. As both low-level semantics of lyrics and our emotional responses can be encoded in words, we propose a simplified cognitive approach based on LSA latent semantic analysis. Modeling how we perceive the emotional content of song lyrics, the multiple contexts in which words occur are `bottom-up' projected as vectors into a semantic space of reduced dimensionality. While patterns of emotional categorization emerge by defining term vector distances to affective adjectives, which as `top-down' labels constrain the latent semantics, according to a psychological plane framed by the dimensions of valence and arousal. | | Keywords | cognitive components, latent semantics, emotions | | Type | Conference paper [With referee] | | Conference | ACM Recommender Systems, Workshop on music recommendation and discovery | | Year | 2010 Month July | | Electronic version(s) | [pdf] | | BibTeX data | [bibtex] | | IMM Group(s) | Intelligent Signal Processing |
|