Searcharxiv⌕ Search

arXiv subjects

Jimena Royo-Letelier

Publications and source records attributed to Jimena Royo-Letelier.

13 recordsLinked to original sources
arXiv2026

Mapping the political landscape from data traces: multidimensional opinions of users, politicians and media outlets on X

Studying political activity on social media often requires defining and measuring political stances of users or content. Relevant examples include the study of opinion polarization, or the study of political diversity in online content diets. While many research designs rely on operationalizations best suited for the US setting, few allow addressing more general political systems, in which users and media outlets might exhibit stances on multiple ideology and issue dimensions, going beyond traditional Liberal-Conservative or Left-Right scales. To advance the study of more general online ecosystems, we present a dataset pertaining to a population of X/Twitter users, parliamentarians, and media outlets embedded in a political space spanned by dimensions measuring attitudes towards immigration, the EU, liberal values, elites and institutions, nationalism and the environment, in addition to left-right and liberal-conservative scales. We include indicators of individual activity and popularity: mean number of posts per day, number of followers, and number of followees. We provide several benchmarks validating the positions of these entities and discuss several applications for this dataset.

cs.SI↗
arXiv2019

Improving Collaborative Metric Learning with Efficient Negative Sampling

Distance metric learning based on triplet loss has been applied with success in a wide range of applications such as face recognition, image retrieval, speaker change detection and recently recommendation with the CML model. However, as we show in this article, CML requires large batches to work reasonably well because of a too simplistic uniform negative sampling strategy for selecting triplets. Due to memory limitations, this makes it difficult to scale in high-dimensional scenarios. To alleviate this problem, we propose here a 2-stage negative sampling strategy which finds triplets that are highly informative for learning. Our strategy allows CML to work effectively in terms of accuracy and popularity bias, even when the batch size is an order of magnitude smaller than what would be needed with the default uniform sampling. We demonstrate the suitability of the proposed strategy for recommendation and exhibit consistent positive results across various datasets.

cs.IR↗
arXiv2019

Singing voice separation: a study on training data

In the recent years, singing voice separation systems showed increased performance due to the use of supervised training. The design of training datasets is known as a crucial factor in the performance of such systems. We investigate on how the characteristics of the training dataset impacts the separation performances of state-of-the-art singing voice separation algorithms. We show that the separation quality and diversity are two important and complementary assets of a good training dataset. We also provide insights on possible transforms to perform data augmentation for this task.

cs.SD↗
arXiv2018

Disambiguating Music Artists at Scale with Audio Metric Learning

We address the problem of disambiguating large scale catalogs through the definition of an unknown artist clustering task. We explore the use of metric learning techniques to learn artist embeddings directly from audio, and using a dedicated homonym artists dataset, we compare our method with a recent approach that learn similar embeddings using artist classifiers. While both systems have the ability to disambiguate unknown artists relying exclusively on audio, we show that our system is more suitable in the case when enough audio data is available for each artist in the train dataset. We also propose a new negative sampling method for metric learning that takes advantage of side information such as music genre during the learning phase and shows promising results for the artist clustering task.

cs.IR↗
arXiv2018

Audio Based Disambiguation Of Music Genre Tags

In this paper, we propose to infer music genre embeddings from audio datasets carrying semantic information about genres. We show that such embeddings can be used for disambiguating genre tags (identification of different labels for the same genre, tag translation from a tag system to another, inference of hierarchical taxonomies on these genre tags). These embeddings are built by training a deep convolutional neural network genre classifier with large audio datasets annotated with a flat tag system. We show empirically that they makes it possible to retrieve the original taxonomy of a tag system, spot duplicates tags and translate tags from a tag system to another.

cs.IR↗
arXiv2018

Music Mood Detection Based On Audio And Lyrics With Deep Neural Net

We consider the task of multimodal music mood prediction based on the audio signal and the lyrics of a track. We reproduce the implementation of traditional feature engineering based approaches and propose a new model based on deep learning. We compare the performance of both approaches on a database containing 18,000 tracks with associated valence and arousal values and show that our approach outperforms classical models on the arousal detection task, and that both approaches perform equally on the valence prediction task. We also compare the a posteriori fusion with fusion of modalities optimized simultaneously with each unimodal model, and observe a significant improvement of valence prediction. We release part of our database for comparison purposes.

cs.IR↗
arXiv2018

Word2Vec applied to Recommendation: Hyperparameters Matter

Skip-gram with negative sampling, a popular variant of Word2vec originally designed and tuned to create word embeddings for Natural Language Processing, has been used to create item embeddings with successful applications in recommendation. While these fields do not share the same type of data, neither evaluate on the same tasks, recommendation applications tend to use the same already tuned hyperparameters values, even if optimal hyperparameters values are often known to be data and task dependent. We thus investigate the marginal importance of each hyperparameter in a recommendation setting through large hyperparameter grid searches on various datasets. Results reveal that optimizing neglected hyperparameters, namely negative sampling distribution, number of epochs, subsampling parameter and window-size, significantly improves performance on a recommendation task, and can increase it by an order of magnitude. Importantly, we find that optimal hyperparameters configurations for Natural Language Processing tasks and Recommendation tasks are noticeably different.

cs.IR↗
arXiv2014

Sharp interface limit for two components Bose-Einstein condensates

We study a double Cahn-Hilliard type functional related to the Gross-Pitaevskii energy of two-components Bose-Einstein condensates. In the case of large but same order intercomponent and intracomponent coupling strengths, we prove $Γ$-convergence to a perimeter minimisation functional with an inhomogeneous surface tension. We study the asymptotic behavior of the surface tension as the ratio between the intercomponent and intracomponent coupling strengths becomes very small or very large. We obtain as a consequence, symmetry breaking of the minimisers for the harmonic potential when this radio is sufficiently large.

math.AP↗
arXiv2014

On the low lying spectrum of the magnetic Schrödinger operator with kagome periodicity

We study in a semiclassical regime a two-dimensional magnetic periodic Schrödinger operator. We first review some results for the square (Harper), triangular and hexagonal (case of the graphene) lattices. Then we study the case considered by Hou when the periodicity is given by a kagome lattice. We reduce the problem to the study of discrete and pseudodifferential operators and obtain pictures similar to Hofstadter's butterfly. We prove the existence of flat bands, which do not occur in the three previous cases.

math.AP↗
arXiv2013

A minimal interface problem arising from a two component Bose Einstein condensate via $\G$-convergence

We consider the energy modeling a two component Bose-Einstein condensate in the limit of strong coupling and strong segregation. We prove the $Γ$-convergence to a perimeter minimization problem, with a weight given by the density of the condensate. In the case of equal mass for the two components, this leads to symmetry breaking for the ground state. The proof relies on a new formulation of the problem in terms of the total density and spin functions, which turns the energy into the sum of two weighted Cahn-Hilliard energies. Then, we use techniques coming from geometric measure theory to construct upper and lower bounds. In particular, we make use of the slicing technique introduced in Ambrosio-Tororelli (CPAM, 1990).

math.AP↗
arXiv2012

Segregation and symmetry breaking of strongly coupled two-component Bose-Einstein condensates in a harmonic trap

We study ground states of two-component condensates in a harmonic trap. We prove that in the strongly coupled and weakly interacting regime, the two components segregate while a symmetry breaking occurs. More precisely, we show that when the intercomponent coupling strength is very large and both intracomponent coupling strengths are small, each component is close to the positive or the negative part of a second eigenfunction of the harmonic oscillator in $\R^2$. As a result, the supports of the components approach complementary half-spaces, and they are not radially symmetric.

math-ph↗
arXiv2010

Non existence of vortices in the small density region of a condensate

In this paper, we answer a question raised by Len Pitaevskii and prove that the ground state of the Gross-Pitaevskii energy describing a Bose Einstein condensate at low rotation does not have vortices in the low density region. Therefore, the first ground state with vortices has its vortices in the bulk. This is obtained by proving that for small rotational velocities, the ground state is multiple of the ground state with zero rotation. We rely on sharp bounds of the decay of the wave function combined with weighted jacobian estimates.

math-ph↗