An official website of the United States government
Every artist has a creative process that draws inspiration from previous artists and their works. Today, “inspiration” has been automated by generative music models. The black box nature of these models obscures the identity of the works that influence their creative output. As a result, users may inadvertently appropriate, misuse, or copy existing artists’ works. We establish a replicable methodology to systematically identify similar pieces of music audio in a manner that is useful for understanding training data attribution. A key aspect of our approach is to harness an effective music audio similarity measure. We compare the effect of applying CLMR [50] and CLAP [55] embeddings to similarity measurement in a set of 5 million audio clips used to train VampNet [24], a recent open source generative music model. We validate this approach with a human listening study. We also explore the effect that modifications of an audio example (e.g., pitch shifting, time stretching, background noise) have on similarity measurements. This work is foundational to incorporating automated influence attribution into generative modeling, which promises to let model creators and users move from ignorant appropriation to informed creation. Audio samples that accompany this paper are available at https://tinyurl.com/exploring-musical-roots.
more »
« less
Robust inference in deconvolution
Kotlarski's identity has been widely used in applied economic research based on repeated‐measurement or panel models with latent variables. However, how to conduct inference for these models has been an open question for two decades. This paper addresses this open problem by constructing a novel confidence band for the density function of a latent variable in repeated measurement error model. The confidence band builds on our finding that we can rewrite Kotlarski's identity as a system of linear moment restrictions. Our approach is robust in that we do not require the completeness. The confidence band controls the asymptotic size uniformly over a class of data generating processes, and it is consistent against all fixed alternatives. Simulation studies support our theoretical results.
more »
« less
- PAR ID:
- 10222883
- Date Published:
- Journal Name:
- Quantitative Economics
- Volume:
- 12
- Issue:
- 1
- ISSN:
- 1759-7323
- Page Range / eLocation ID:
- 109 to 142
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Latent diffusion models have been demonstrated to generate high-quality images, while offering efficiency in model training compared to diffusion models operating in the pixel space. However, incorporating latent diffusion models to solve inverse problems remains a challenging problem due to the nonlinearity of the encoder and decoder. To address these issues, we propose ReSample, an algorithm that can solve general inverse problems with pre-trained latent diffusion models. Our algorithm incorporates data consistency by solving an optimization problem during the reverse sampling process, a concept that we term as hard data consistency. Upon solving this optimization problem, we propose a novel resampling scheme to map the measurement-consistent sample back onto the noisy data manifold and theoretically demonstrate its benefits. Lastly, we apply our algorithm to solve a wide range of linear and nonlinear inverse problems in both natural and medical images, demonstrating that our approach outperforms existing state-of-the-art approaches, including those based on pixel-space diffusion models.more » « less
-
Focusing on developing maker-identities, especially for historically marginalized students in the computational field, can empower them to recognize and take ownership of their space in the field. Drawing from identity related literature in maker and computing related fields we identified seven factors of maker-identity - interest and motivation, competence and performance, confidence and self-efficacy, recognition, utility value and meaningfulness, perceptions of community, and external factors. Using this, we analyzed semi-structured interviews of students who participated in our summer makerspace camp to understand how these identity factors manifested in their reflections of the camp. We tie back our findings of positive impacts on maker-identity structures to the design structures of our makerspace such as co-design of the space, use-modify-create strategies, and open-ended design projects.more » « less
-
We propose a novel unsupervised generative model that learns to disentangle object identity from other low-level aspects in class-imbalanced data. We first investigate the issues surrounding the assumptions about uniformity made by InfoGAN [10], and demonstrate its ineffectiveness to properly disentangle object identity in imbalanced data. Our key idea is to make the discovery of the discrete latent factor of variation invariant to identity-preserving transformations in real images, and use that as a signal to learn the appropriate latent distribution representing object identity. Experiments on both artificial (MNIST, 3D cars, 3D chairs, ShapeNet) and real-world (YouTube-Faces) imbalanced datasets demonstrate the effectiveness of our method in disentangling object identity as a latent factor of variation.more » « less
-
Abstract Identity development frameworks provide insight into why and to what extent individuals engage in STEM‐related activities. While studies of “STEM identity” often build off previously validated disciplinary and/or science identity frameworks, quantitative analyses of constructs that specifically measure STEM identity and its antecedents are scarce, making it challenging for researchers or practitioners to apply a measurement‐based perspective of participation in opportunities billed as “STEM.” In this study, we tested two expanded structural equation models of STEM identity development, building off extensions of science and disciplinary‐identity frameworks, that incorporated additional factors relevant to identity development: gender, ethnicity, home science support, parental education, and experiencing science talk in the home. Our models test theorized relationships between interest, sense of recognition, performance‐competence, and identity in the context of STEM with undergraduate students (N = 522) enrolled in introductory STEM courses at a Hispanic Serving Institution. Our findings support our measurement of STEM identity and its indicators, providing researchers with a predictive model associated with academic intentions across disciplinary domains in STEM. Further, our expanded model (i.e., Model I+) indicates significant contributions of participant gender, which has a larger indirect effect on STEM identity (β = 0.50) than the direct effect of STEM interest (β = 0.29), and of home support in relation to performance‐competence in academic contexts. Our model also posits a significant contribution of family science talk to sense of recognition as a STEM person, expanding our understandings of the important role of the home environment while challenging prior conceptions of science capital and habitus. We situate our results within a broader discussion regarding the validity of “STEM identity” as a concept and construct in the context of communities often marginalized in STEM fields.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3982/QE1643
