Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Computational Modeling of Artistic Inspiration: A Framework for Predicting Aesthetic Preferences in Lyrical Lines Using Linguistic and Stylistic Features

Gaurav Sahu, Olga Vechtomova

TL;DR

This work proposes a novel framework for computationally modeling artistic preferences in different individuals through key linguistic and stylistic properties, with a focus on lyrical content, and introduces a dataset of annotated lyric lines to facilitate the evaluation of the framework.

Abstract

Artistic inspiration remains one of the least understood aspects of the creative process. It plays a crucial role in producing works that resonate deeply with audiences, but the complexity and unpredictability of aesthetic stimuli that evoke inspiration have eluded systematic study. This work proposes a novel framework for computationally modeling artistic preferences in different individuals through key linguistic and stylistic properties, with a focus on lyrical content. In addition to the framework, we introduce \textit{EvocativeLines}, a dataset of annotated lyric lines, categorized as either "inspiring" or "not inspiring," to facilitate the evaluation of our framework across diverse preference profiles. Our computational model leverages the proposed linguistic and poetic features and applies a calibration network on top of it to accurately forecast artistic preferences among different creative individuals. Our experiments demonstrate that our framework outperforms an out-of-the-box LLaMA-3-70b, a state-of-the-art open-source language model, by nearly 18 points. Overall, this work contributes an interpretable and flexible framework that can be adapted to analyze any type of artistic preferences that are inherently subjective across a wide spectrum of skill levels.

Computational Modeling of Artistic Inspiration: A Framework for Predicting Aesthetic Preferences in Lyrical Lines Using Linguistic and Stylistic Features

TL;DR

This work proposes a novel framework for computationally modeling artistic preferences in different individuals through key linguistic and stylistic properties, with a focus on lyrical content, and introduces a dataset of annotated lyric lines to facilitate the evaluation of the framework.

Abstract

Artistic inspiration remains one of the least understood aspects of the creative process. It plays a crucial role in producing works that resonate deeply with audiences, but the complexity and unpredictability of aesthetic stimuli that evoke inspiration have eluded systematic study. This work proposes a novel framework for computationally modeling artistic preferences in different individuals through key linguistic and stylistic properties, with a focus on lyrical content. In addition to the framework, we introduce \textit{EvocativeLines}, a dataset of annotated lyric lines, categorized as either "inspiring" or "not inspiring," to facilitate the evaluation of our framework across diverse preference profiles. Our computational model leverages the proposed linguistic and poetic features and applies a calibration network on top of it to accurately forecast artistic preferences among different creative individuals. Our experiments demonstrate that our framework outperforms an out-of-the-box LLaMA-3-70b, a state-of-the-art open-source language model, by nearly 18 points. Overall, this work contributes an interpretable and flexible framework that can be adapted to analyze any type of artistic preferences that are inherently subjective across a wide spectrum of skill levels.
Paper Structure (33 sections, 21 equations, 5 figures, 6 tables)

This paper contains 33 sections, 21 equations, 5 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. The Proposed Framework
  4. Step 1: Identifying The Key Poetic and Stylistic Characteristics
  5. Poetic/Stylistic Characteristics.
  6. Poetic Imagery
  7. Word Energy
  8. Level of Abstraction
  9. Valence
  10. Linguistic and Statistical Characteristics.
  11. Surprisal
  12. Contextual Entropy
  13. Normalized Pointwise Mutual Information (NPMI)
  14. Banality
  15. Step 2: Training the Calibration Network
  16. ...and 18 more sections

Figures (5)

  • Figure 1: Inter-annotator agreement on the EvocativeLines (small) dataset.
  • Figure 2: Radar charts showing the 10 different preference profiles of annotators in the EvocativeLines dataset. Blue webs denotes the web for positive lines, and red webs denotes the negative lines.
  • Figure 3: Examples of lyric lines rated as inspiring (blue) and not inspiring (red) by annotator $\mathcal{A}_1$.
  • Figure 4: Guidelines presented to the eight human labelers.
  • Figure 5: Prompt used to measure poetic imagery of a lyric line.