Discovering Interpretable Concepts in Large Generative Music Models

• URL: http://arxiv.org/abs/2505.18186v1
• Date: Sun, 18 May 2025 19:44:20 GMT
• Title: Discovering Interpretable Concepts in Large Generative Music Models
• Authors: Nikhil Singh, Manuel Cherep, Pattie Maes,
• Abstract summary: We introduce a method to discover musical concepts using sparse autoencoders (SAEs)<n>We evaluate this approach by extracting a large set of features and producing an automatic labeling and evaluation pipeline.<n>Our results reveal both familiar musical concepts and counterintuitive patterns that lack clear counterparts in existing theories or natural language altogether.
• Score: 23.427995912149385
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The fidelity with which neural networks can now generate content such as music presents a scientific opportunity: these systems appear to have learned implicit theories of the structure of such content through statistical learning alone. This could offer a novel lens on theories of human-generated media. Where these representations align with traditional constructs (e.g. chord progressions in music), they demonstrate how these can be inferred from statistical regularities. Where they diverge, they highlight potential limits in our theoretical frameworks -- patterns that we may have overlooked but that nonetheless hold significant explanatory power. In this paper, we focus on the specific case of music generators. We introduce a method to discover musical concepts using sparse autoencoders (SAEs), extracting interpretable features from the residual stream activations of a transformer model. We evaluate this approach by extracting a large set of features and producing an automatic labeling and evaluation pipeline for them. Our results reveal both familiar musical concepts and counterintuitive patterns that lack clear counterparts in existing theories or natural language altogether. Beyond improving model transparency, our work provides a new empirical tool that might help discover organizing principles in ways that have eluded traditional methods of analysis and synthesis.

Related papers

• Concept-Guided Interpretability via Neural Chunking [54.73787666584143]
  We show that neural networks exhibit patterns in their raw population activity that mirror regularities in the training data.<n>We propose three methods to extract these emerging entities, complementing each other based on label availability and dimensionality.<n>Our work points to a new direction for interpretability, one that harnesses both cognitive principles and the structure of naturalistic data.
  arXiv  Detail & Related papers  (2025-05-16T13:49:43Z)
• I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data? [76.15163242945813]
  Large language models (LLMs) have led many to conclude that they exhibit a form of intelligence.<n>We introduce a novel generative model that generates tokens on the basis of human-interpretable concepts represented as latent discrete variables.
  arXiv  Detail & Related papers  (2025-03-12T01:21:17Z)
• Deep Learning Through A Telescoping Lens: A Simple Model Provides Empirical Insights On Grokking, Gradient Boosting & Beyond [61.18736646013446]
  In pursuit of a deeper understanding of its surprising behaviors, we investigate the utility of a simple yet accurate model of a trained neural network. Across three case studies, we illustrate how it can be applied to derive new empirical insights on a diverse range of prominent phenomena.
  arXiv  Detail & Related papers  (2024-10-31T22:54:34Z)
• Distilling Symbolic Priors for Concept Learning into Neural Networks [9.915299875869046]
  We show that inductive biases can be instantiated in artificial neural networks by distilling a prior distribution from a symbolic Bayesian model via meta-learning. We use this approach to create a neural network with an inductive bias towards concepts expressed as short logical formulas.
  arXiv  Detail & Related papers  (2024-02-10T20:06:26Z)
• Unraveling Feature Extraction Mechanisms in Neural Networks [10.13842157577026]
  We propose a theoretical approach based on Neural Tangent Kernels (NTKs) to investigate such mechanisms. We reveal how these models leverage statistical features during gradient descent and how they are integrated into final decisions. We find that while self-attention and CNN models may exhibit limitations in learning n-grams, multiplication-based models seem to excel in this area.
  arXiv  Detail & Related papers  (2023-10-25T04:22:40Z)
• Deep Generative Models of Music Expectation [2.900810893770134]
  We propose to use modern deep probabilistic generative models in the form of a Diffusion Model to compute an approximate likelihood of a musical input sequence. Unlike prior work, such a generative model parameterized by deep neural networks is able to learn complex non-linear features directly from a training set itself. We show that pre-trained diffusion models indeed yield musical surprisal values which exhibit a negative quadratic relationship with measured subject 'liking' ratings.
  arXiv  Detail & Related papers  (2023-10-05T12:25:39Z)
• A Recursive Bateson-Inspired Model for the Generation of Semantic Formal Concepts from Spatial Sensory Data [77.34726150561087]
  This paper presents a new symbolic-only method for the generation of hierarchical concept structures from complex sensory data. The approach is based on Bateson's notion of difference as the key to the genesis of an idea or a concept. The model is able to produce fairly rich yet human-readable conceptual representations without training.
  arXiv  Detail & Related papers  (2023-07-16T15:59:13Z)
• Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker [72.09076317574238]
  ToM is a plug-and-play approach to investigate the belief states of characters in reading comprehension. We show that ToM enhances off-the-shelf neural network theory mind in a zero-order setting while showing robust out-of-distribution performance compared to supervised baselines.
  arXiv  Detail & Related papers  (2023-06-01T17:24:35Z)
• Learning Unsupervised Hierarchies of Audio Concepts [13.400413055847084]
  In computer vision, concept learning was proposed to adjust explanations to the right abstraction level. In this paper, we adapt concept learning to the realm of music, with its particularities. We propose a method to learn numerous music concepts from audio and then automatically hierarchise them to expose their mutual relationships.
  arXiv  Detail & Related papers  (2022-07-21T16:34:31Z)
• Sequence Generation using Deep Recurrent Networks and Embeddings: A study case in music [69.2737664640826]
  This paper evaluates different types of memory mechanisms (memory cells) and analyses their performance in the field of music composition. A set of quantitative metrics is presented to evaluate the performance of the proposed architecture automatically.
  arXiv  Detail & Related papers  (2020-12-02T14:19:19Z)
• From Artificial Neural Networks to Deep Learning for Music Generation -- History, Concepts and Trends [0.0]
  This paper provides a tutorial on music generation based on deep learning techniques. It analyzes some early works from the late 1980s using artificial neural networks for music generation.
  arXiv  Detail & Related papers  (2020-04-07T00:33:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.