Human and Machine Intelligence: a Music Information Retrieval perspective

I had the honor to be invited as keynote speaker at the International Conference on Computational Creativity 2020. The conference took place in a virtual format given the health emergency situation. In my talk, I discussed on the different motivations for music information retrieval, the paradigm shift from knowledge-driven to data-driven systems, the main challenges of this field of research and its impact on society.

It was an interesting experience to give a virtual keynote. First, it was a lot of work to prepare for the keynote, including recording the video properly at home. In addition, I missed the experience of seeing people’s faces and expression when listening to your talk. This visual feedback is really important to see how people react to the different parts of your talk. And I missed the travel and visit to beautiful Coimbra. Virtual events also have positive sides, as the confort of speaking at home, accessibility for anyone in the world to follow on streaming, and all energy saved by people not travelling to sites.

You can watch it in youtube! Any comment or feedback is welcome.

Leave a comment

Filed under keynotes

Música y tecnología para el confinamiento

Gracias a la sugerencia de mi tocayo Emilio, he empezado ésta entrada con algunas propuestas de tecnología musical para ésta época en que nos toca estar confinados o aislados de nuestra familia y amigos. La iré actualizando con ideas y propuestas y si queréis contribuir, por favor contactadme o comentar en éste post.

Música clásica

Muchos de los contenidos que éstos días están disponibles vienen de óperas y otros espectáculos de música clásica.

La Metropolitan Opera ofrece acceso libre a sus óperas a través de su web. La semana pasada la dedicaron a Wagner https://www.metopera.org/user-information/nightly-met-opera-streams/ 

La Berliner Philharmoniker también proporciona sus conciertos de manera gratuita desde el ordenador, tablet o móvil en su  Digital Concert Hall, que se puede descargar desde, por ejemplo el Appstore, y acceder gratis durante 30 días con el código BERLINPHIL.

El Teatro Real permite en éstos días el acceso gratuito a su plataforma de vídeo “My Opera Player”, para poder disfrutar de todos los títulos de ópera, danza y conciertos que ya ofrece en su catálogo y de los que se están incorporando estos días. Para acceder hay que registrarse en la página web y utilizar el código OperaEnCasa. Además de las obras del teatro real hay espectáculos de diferentes teatros en abierto desde ésta web.

El Gran Teatre del Liceu también ha puesto a disposición en su web una serie de títulos y éxitos de temporadas recientes a la web #theliceuathome.

    • Gracias a la colaboración con el Teatro Real además las óperas NormaRigoletto estarán disponibles de forma gratuita a la plataforma My Opera Player con el código OperaEnCasa.
    • En la web del canal ARTE podemos disfrutar de forma gratuita de su producción en alta tecnología de la ópera  Turandot de Puccini que abrió la temporada 2019-20 del Liceu.

Auditori de Barcelona ha habilitado una serie de contenidos digitales en abierto en su web,  en su página web titulada l’Auditori en tu casa. Se incluyen una serie de libros que tienen la música como temática, recomendaciones de series y películas, un directo en su Facebook cada día a las 20 con los conciertos de diferentes salas europeas, una lista de reproducción de Spotify, un ejercicio didáctico de regulación de emociones con la música, y una serie de conciertos como uno que hizo la Orquesta Sinfónica de Barcelona y Nacional de Cataluña (OBC) en playa de Barcelona donde podemos escuchar el murmullo del mar.

La Ópera de Viena está publicando cada día una de las óperas de su archivo. En su calendario podemos encontrar diversas óperas programadas hasta final del mes de junio.

Para niños

El Gran Teatre del Liceu también ha habilitado en su canal de  Youtube varios de los espectáculos con más éxito de El Petit Liceu pensados para los pequeños de la casa, que están en catalán: La petita flauta màgicaLa ventafocs y Pere i el Llop. 

Música pop

    • Podéis aprovechar éstos días para crear vuestra propia banda sonora a través de nuestra web. 
    • Además, hemos elaborado las listas que se han ido creando en el proyecto para construir una lista con las 157 canciones, unas 7 horas de música, ordenadas de mayor éxito a menor,. Podría ser algo que las personas mayores que están confinadas éstos días pueden agradecer. Incluye éxitos de Antonio Molina, Marisol, Paloma Sanbasilio o Rocío Jurado. La lista está pública en éste enlace: http://bit.ly/bandasonoravital 

Y aquí en España se han creado festivales virtuales como el Festival me quedo en casa  o la Cuarentena fest donde muchos artistas independientes, a través de sus perfiles de instagram están haciendo directos de unos 30 minutos como si de un concierto acústico privado se tratase. Los horarios están en las páginas que hemos enlazado.

Flamenco, rock, jazz

Durante éstos días también podéis aprender a cantar flamenco gracias a la tecnología y a nuestro curso online CanteFlamencoTech, disponible en la plataforma Miriadax. Está pensado para una duración de 4 semanas y se inició el pasado día 16 pero los contenidos están abiertos aún para poder seguir el curso.

El festival de jazz de Montreaux ofrece acceso abierto a 50 de sus conciertos mediante su plataforma de streaming. Éstos incluyen interpretaciones de  Ray Charles, Wu-Tang Clan, Johnny Cash, Nina Simone, Marvin Gaye, Deep Purple, Carlos Santana, y muchos más artistas. La información (en inglés) de cómo acceder se puede encontrar en su web, que consiste en entrar en la plataforma y utilizar un código de descuento.

 

Acknowledgements: gracias a las aportaciones de Emilio Gómez-González, Jesús Crespo-Cuaresma

Leave a comment

Filed under personal, research

Research on child-robot interaction

In the context of the HUMAINT (Human behaviour and machine intelligence) project, we research on the impact that social robots have on children. In this context, I have had the chance to carry out my first research on the amazing field of child-robot interaction, thanks to the collaboration with Vicky Charisi, Luis Merino and their lab at Universidad Pablo Olavide and Honda Research Institute Japan.

Running a user study with children and robots is very challenging from a technical perspective, and  analysing their data is challenging as well.  We just published in frontiers the result of our first study, where we experimented with two strategies for child-robot interaction in a problem solving task: turn taking and child-initiated interaction, and we showed the need for this voluntary interaction. You can check the details below. It is amazing to learn and contribute to research on this topic!

Child-Robot Collaborative Problem-Solving and the Importance of Child’s Voluntary Interaction: A Developmental Perspective

Vicky Charisi, Emilia Gomez, Gonzalo Mier, Luis Merino and Randy Gomez

Abstract: The emergence and development of cognitive strategies for the transition from exploratory actions towards intentional problem-solving in children is a key question for the understanding of the development of human cognition. Researchers in developmental psychology have studied cognitive strategies and have highlighted the catalytic role of the social environment. However, it is not yet adequately understood how this capacity emerges and develops in biological systems when they perform a problem-solving task in collaboration with a robotic social agent. This paper presents an empirical study in a human-robot interaction (HRI) setting which investigates children’s problem-solving from a developmental perspective. In order to theoretically conceptualize children’s developmental process of problem-solving in HRI context, we use principles based on the intuitive theory and we take into consideration existing research on executive functions with a focus on inhibitory control. We considered the paradigm of the Tower of Hanoi and we conducted an HRI behavioral experiment to evaluate task performance. We designed two types of robot interventions, “voluntary” and “turn-taking”—manipulating exclusively the timing of the intervention. Our results indicate that the children who participated in the voluntary interaction setting showed a better performance in the problem solving activity during the evaluation session despite their large variability in the frequency of self-initiated interactions with the robot. Additionally, we present a detailed description of the problem-solving trajectory for a representative single case-study, which reveals specific developmental patterns in the context of the specific task. Implications and future work are discussed regarding the development of intelligent robotic systems that allow child-initiated interaction as well as targeted and not constant robot interventions.

 

Leave a comment

Filed under publications, research

My last day as ISMIR president

This is my last day as ISMIR president. I am very proud of how our community has managed to establish ISMIR as a leading forum for music information research. I joined the board in 2014, became president-elect in 2016 and president in 2018. My main motivations have always been to contribute to strengthen scientific quality, inclusive spirit and diversity. During these years, I have witnessed the origins and impulse of TISMIR and WiMIR, the consolidation of ISMIR location alternating between Europe, America and Asia, and the evolution of the MIR field including the involvement of industry in our community, changes of on topics and MIR-related disciplines, as reflected in its 20th anniversary edition.

But, among all, I have had the privilege to meet great researchers and people in a community I now considered a bit as my family. I want to thank all excellent board members and colleagues I had the chance to work with. I do not have enough space to express my gratefulness.

Now it is my time to take a break from scientific service, which is sometimes exhausting and no so well recognised. However, I will keep my compromise as TISMIR co-editor in chief, WiMIR mentor and PC member, and hopefully as an ISMIR author for many years. I am sure the next president and board will do a great job and I wish them all the best to face the main challenge I foresee: how to keep ISMIR spirit (single track conference), scientific insights, inclusion and diversity in a deep learning world.

 

Leave a comment

Filed under research

Journal paper on AI and Music: Open Questions of Copyright Law and Engineering Praxis

I am very happy to share with you the publication of a truly interdisciplinary study on the impact of AI on music, including considerations from copyright and engineering praxis. It has been an amazing experience to collaborate with scholars in the field of creative practices, engineering and law, and I hope the paper will serve to start discussing some relevant aspects related to the use of AI in music production.

Abstract

The application of artificial intelligence (AI) to music stretches back many decades, and presents numerous unique opportunities for a variety of uses, such as the recommendation of recorded music from massive commercial archives, or the (semi-)automated creation of music. Due to unparalleled access to music data and effective learning algorithms running on high-powered computational hardware, AI is now producing surprising outcomes in a domain fully entrenched in human creativity—not to mention a revenue source around the globe. These developments call for a close inspection of what is occurring, and consideration of how it is changing and can change our relationship with music for better and for worse. This article looks at AI applied to music from two perspectives: copyright law and engineering praxis. It grounds its discussion in the development and use of a specific application of AI in music creation, which raises further and unanticipated questions. Most of the questions collected in this article are open as their answers are not yet clear at this time, but they are nonetheless important to consider as AI technologies develop and are applied more widely to music, not to mention other domains centred on human creativity.

Keywords: artificial intelligence; music; copyright; engineering; ethics

 

Paper available in open access at Arts journal.

Leave a comment

Filed under publications, research

OpenBMAT: a new open dataset for music detection with loudness annotations

Last week we announced the publication of OpenBMAT, an open dataset for the tasks of music detection and relative music loudness estimation. The dataset contains 27.4 hours of audio from 8 different TV program types at 4 different countries, cross-annotated by 3 people using 6 different classes. It has been published as a dataset paper at Transaction of the International Society for Music Information Retrieval, the open journal of ISMIR. This research has been carried out as a collaboration between the MTG and BMAT in the context of the industrial Doctorates program of the Catalan Government.

For more information you can read the related news at MTG web site: https://www.upf.edu/web/mtg/home/-/asset_publisher/sWCQhjdDLWwE/content/id/227864284/maximized#.XXZ_IZMzab8

 

Leave a comment

Filed under publications, research, Uncategorized

Help us in our study on perceived emotions of pop/rock music by English, Spanish, German, and Mandarin speakers

En castellano debajo! 

Juan Sebastian Gómez-Cañón, one of the PhD students working with me at the Music Information Research Lab of the MTG, is leading a study on the relation between the emotions perceived in pop and rock music and the language the listener speaks. To participate, please visit the following links depending on your mother tongue: English, Spanish, German, and Mandarin. In case that none of these languages is native to you, you can still participate by filling the English version (it takes around 20-25 minutes to complete it). Please follow the instructions carefully to guarantee your comprehension, correctness, and overall enjoyment! You can stop at any time and continue later as long as you keep the volume at the same level.

At the end of this survey and as a small thank you for your time, we will provide your Music Sophistication Index, which measures your ability to engage with music, as defined by Müllensiefen et al. (2014). The results of this research will also be available to you in our project website.

Thank you!

——

Juan Gómez-Canón, con el  cual trabajo en el laboratorio de Music Information Research del MTG (Universitat Pompeu Fabra, Barcelona) está llevando a cabo un estudio sobre la relación entre las emociones percibidas en música pop/rock y el lenguaje. Agradeceríamos mucho vuestra participación rellenando un formulario que requiere entre 20-25 minutos y que se basa, fundamentalmente, en escuchar música e indicar la emoción percibida en cada caso. El formulario está disponible en Inglés, Español, Alemán y Mandarín (en caso que vuestra lengua materna no sea ninguna de esas podéis participar usando la versión en inglés). Es importante seguir las instrucciones cuidadosamente para garantizar la correcta comprensión de los ítems planteados. Podéis parar en cualquier momento y continuar después, siempre y cuando se mantenga el mismo volumen al escuchar la música .

Al final de esta encuesta y como una pequeña muestra de agradecimiento por vuestro tiempo, os entregaremos vuestro Music Sophistication Index, que mide la habilidad de involucrarse con la música, tal y como se definió en Müllensiefen et al. (2014). Los resultados de esta investigación también estarán disponibles para la página web del proyecto.

 

¡Muchas gracias!

v—-

Leave a comment

Filed under research

Women in AI: mitigating the gender bias

Last Friday March 8th I was invited to speak at a lunch event of the European Commission intended to provide a scientific perspective to the challenges of gender equality. I gave a talk titled “Women in Artificial Intelligence: mitigating the gender bias”, that is summarized here.

In this context today my colleague Ana Freire and I are launching the divinAI initative to monitor the presence of women in AI events. Please come to our HACKFEST event in Barcelona in June 1st! 

 

Hackathon.png

Leave a comment

Filed under keynotes, research

El papel que juega la intuición en la interpretación musical

Una investigación realizada por Álvaro Sarasúa, Julián Urban y una servidora en el Grupo de Investigación en Tecnología Musical, dentro del proyecto europeo Phenicx,  describe las ventajas de usabilidad de un sistema personalizado que aprovecha la intuición de los usuarios al interpretar música.

 

Intuición que por cierto no tienen aún los sistemas automáticos a los que tenemos que entrenar desde 0.

Podéis leer la noticia en la web de la Universitat Pompeu Fabra. 

 

Comments Off on El papel que juega la intuición en la interpretación musical

27/02/2019 · 14:53

El dilema de la tecnología ética – The dilemma of ethical technologies

Article in Spanish about ethics of technology written by Esther Paniagua in the context of the HUMAINT winter school that took place this month at JRC Seville. The article is also available online.

Leave a comment

Filed under press

A new paper on Frontiers journal on Music Conducting

I am very happy to publish this work with Alvaro Sarasúa and Julián Urbano on Frontiers in Digital Humanities about Music Conducting.

The paper, titled “Mapping by Observation: Building a User-Tailored Conducting System From Spontaneous Movements” presents a music interaction system based on the conductor-orchestra metaphor, where the orchestra is considered as an instrument controlled by the movements of the conductor. In the system we proposed the user can control tempo and dynamics and it adapts its mapping to the user by observing spontaneous conducting movements on top of a fixed music. In this respect, we analyze the tendency of people to anticipate or fall behind the beat and the gestures mapped to loudness. The system was evaluated with 24 participants in a discover-by-playing scenario.
Our work was developed in the context of the PHENICX and CASAS research projects and opens interesting directions for creating more intuitive and expressive DMIs, particularly in public installations.
You can access the open publication here.
fdigh-06-00003-g003

Leave a comment

Filed under publications, research

Entrevista en el mundo sobre inteligencia artificial

Entrevista en el mundo por Silvia Moreno sobre la inteligencia artificial y el impacto que está y tendrá en el futuro en las personas

https://www.elmundo.es/papel/lideres/201 8/09/25/5ba81ccb268e3eee488b462d.html   

 

Leave a comment

Filed under press, womenTIC

A los seres humanos siempre nos quedará la creatividad, una máquina no puede crear ¿mito o realidad?

Un artículo muy interesante de Esther Paniagua sobre la creatividad computacional donde comento que un algoritmo de generación de sonido o música puede ser creativo por sí mismo y generar material musical interesante, complementando la creatividad humana. El artículo incluye muchos ejemplos interesantes de arte creado por algoritmos y de opiniones sobre si las máquinas pueden crear. Pues sí, es una realidad!

https://www.xataka.com/robotica-e-ia/a-seres-humanos-siempre-nos-quedara-creatividad-maquina-no-puede-crear-mito-realidad 

Leave a comment

Filed under outreach, press

Saltando como locos de una canción a otra

Así se titula el reportaje de Joseba Elola que se publicó el 29 de Noviembre en el monográfico de música del País semanal, en donde refleja nuestra conversación telefónica donde intenté explicar de manera clara cómo funcionan los sistemas de recomendación musical:

Estos sistemas de recomendación procesan datos editoriales de las canciones como sus títulos o los nombres de los artistas; datos del sonido, como el timbre de la voz, los patrones rítmicos o la escala; e información de los usuarios que han escuchado esa música: si a alguien a quien le gusta Leonard Cohen también le gusta Rufus Wainwright, el algoritmo lo identifica y manda una canción del segundo a aquel que está escuchando al primero. Así lo explica Emilia Gómez, ingeniera de telecomunicaciones especializada en los sistemas de recuperación de la información musical (en inglés, music information retrieval) y presidenta de International Society of Music Information Retrieval (ISMIR), comunidad académica que investiga las tecnologías de recomendación —y con la que colaboran Spotify, Apple o Amazon Music—.

Analizando los instrumentos que suenan, el timbre, la melodía, el ritmo, la estructura de la canción, las voces y el estilo, se confecciona un modelo de los gustos de cada persona, advierte Gómez en conversación telefónica desde Sevilla. “Estos sistemas son cada vez más complejos”, dice, “y las bases de datos, más grandes”. De modo que la tecnología es cada vez más precisa, mejor. 

Esa es la parte buena de la ecuación. La mala, que el predominio de las listas cocinadas por las plataformas acabe por uniformizar lo que escuchamos. “Cuanta más gente hay en las plataformas, mejor funcionan los algoritmos”, manifiesta Gómez. “Mejor, sobre todo, si te gusta lo que a la mayoría. Los que escuchan propuestas raras o minoritarias verán que esa música no se recomienda porque no hay mucha gente que la escuche. Así, se reproduce el sistema que teníamos en la antigüedad, cuando solo un tipo de música se hacía popular, la que te recomendaba la radio”.

No es sólo importante que éstos algoritmos pueden uniformizar  lo que escuchamos, lo cual está relacionado con el denominado efecto “larga cola” en el que siempre se recomienda un pequeño repertorio. Un efecto aún más desconocido es la especialización excesiva del gusto musical que se deriva de la personalización, y que pueden crear éste tipo de algoritmos basados en similitud. En el proyecto HUMAINT estamos estudiando éste efecto, que se conoce como ¨filtro burbuja“, dentro del análisis del impacto de la inteligencia artificial en las personas.

 

Leave a comment

Filed under press

Nuevo MOOC Cante Flamenco Tech

Me alegra poder anunciar un nuevo MOOC abierto y gratuito que empezará el próximo martes 23 de Octubre. Se llama Cante Flamenco Tech, y presenta una introducción al cante flamenco a través de la tecnología.

En éste MOOC he colaborado con la cantaora y profesora Alba Guerrero y con Sonia Rodríguez, y estamos muy orgullosas del resultado.

Es un curso abierto y gratuito, en la plataforma Miríadax, que dura 4 semanas y empieza el próximo martes. ¡Animaros! La inscripción está disponible aquí.

 

 

Leave a comment

Filed under teaching

TISMIR Journal Launch and Call for Papers

I am blogging some news related to a project I have been recently contributing.

It brings us great pleasure to announce the launch of the first issue of TISMIR, the Transactions of the International Society for Music Information Retrievalhttps://transactions.ismir.net/

TISMIR was established to complement the widely cited ISMIR conference proceedings and provide a vehicle for the dissemination of the highest quality and most substantial scientific research in MIR. TISMIR retains the Open Access model of the ISMIR Conference proceedings, providing rapid access, free of charge, to all journal content. In order to encourage reproducibility of the published research papers, we provide facilities for archiving the software and data used in the research. The TISMIR publication model avoids excessive cost to the authors or their institutions, with article charges being less than the ISMIR Conference registration fee.

The first issue contains an editorial introducing the journal, four research papers and one dataset paper:

Editorial: Introducing the Transactions of the International Society for Music Information Retrieval – Simon Dixon,  Emilia Gómez,  Anja Volk

Multimodal Deep Learning for Music Genre Classification – Sergio Oramas,  Francesco Barbieri,  Oriol Nieto,  Xavier Serra

Learning Audio–Sheet Music Correspondences for Cross-Modal Retrieval and Piece Identification – Matthias Dorfer,  Jan Hajič jr.,  Andreas Arzt,  Harald Frostel,  Gerhard Widmer

A New Curated Corpus of Historical Electronic Music: Collation, Data and Research Findings – Nick Collins,  Peter Manning,  Simone Tarsitani

A Case for Reproducibility in MIR: Replication of ‘A Highly Robust Audio Fingerprinting System’ – Joren Six,  Federica Bressan,  Marc Leman

Pop Music Highlighter: Marking the Emotion Keypoints – Yu-Siang Huang,  Szu-Yu Chou,  Yi-Hsuan Yang

Two more papers (one research paper and one overview paper) are in press.

Authors:  We look forward to receiving new submissions to the journal – please see the Call for Papers below.

Best Regards
Simon Dixon, Anja Volk and Emilia Gómez

Editors-in-chief, TISMIR

CALL FOR PAPERS

The ISMIR Board is happy to announce the launch of the Transactions of the International Society for Music Information Retrieval (TISMIR), the open-access journal of our community.

 

TISMIR (http://tismir.ismir.net) publishes novel scientific research in the field of Music Information Retrieval (MIR), an interdisciplinary research area concerned with processing, analysing, organising and accessing music information. We welcome submissions from a wide range of disciplines, including computer science, musicology, cognitive science, library & information science, machine learning, and electrical engineering.

 

TISMIR is established to complement the widely cited ISMIR conference proceedings and provide a vehicle for the dissemination of the highest quality and most substantial scientific research in MIR. TISMIR retains the Open Access model of the ISMIR Conference proceedings, providing rapid access, free of charge, to all journal content. In order to encourage reproducibility of the published research papers, we provide facilities for archiving the software and data used in the research. TISMIR is published in electronic-only format, making it possible to offer very low publication costs to authors’ institutions, while ensuring fully open access content. With this call for papers we invite submissions for the following article types:

Article types

Research articles must describe the outcomes and application of unpublished original research. These should make a substantial contribution to knowledge and understanding in the subject matter and should be supported by relevant experiments.

Overview articles should focus in detail on specific aspects of MIR research. Overview articles will provide a comprehensive review of a broad MIR research problem, a critical evaluation of proposed techniques and/or an analysis of challenges for future research. Papers should critically engage with the relevant body of extant literature.

Datasets should present novel efforts in data gathering and annotation that have a strong potential impact in the way MIR technologies are exploited and evaluated.

 

If the paper extends or combines the authors’ previously published research, it is expected that there is a significant novel contribution in the submission (as a rule of thumb, we would expect at  least 50% of the underlying work – the ideas, concepts, methods, results, analysis and discussion – to be new). In addition, if there is any overlapping textual material, it should be rewritten.

 

Review process

The journal operates a double-blind peer review process.  Review criteria include originality, consideration of previous work, methodology, clarity and reproducibility.

 

Publication frequency

The journal is published online as a continuous volume and issue throughout the year, following an open access policy. Articles are made available as soon as they are ready to ensure that there are no unnecessary delays in getting content publicly available.

 

Editorial team

Editors in Chief

Simon Dixon, School of Electronic Engineering and Computer Science, Queen Mary University of London, United Kingdom

Emilia Gómez, Music Technology Group, Universitat Pompeu Fabra, Barcelona, Spain

Anja Volk, Department of Information and Computing Sciences, Utrecht University, Netherlands

Editorial Board

Juan P. Bello, Department of Music and Performing Arts Professions, & Department of Electrical and Computer Engineering, New York University, United States

Arthur Flexer, Austrian Research Institute for Artificial Intelligence (OFAI), Austria

Fabien Gouyon, Pandora, United States

Xiao Hu, Faculty of Education, Division of Information & Technology Studies, University of Hong Kong

Olivier Lartillot, Department of Musicology, University of Oslo, Norway

Jin Ha Lee, Information School, University of Washington, United States

Meinard Mueller, International Audio Laboratories Erlangen, Germany

Geoffroy Peeters, Sound Analysis/Synthesis Team, UMR STMS IRCAM CNRS, France

Markus Schedl, Department of Computational Perception, Johannes Kepler University Linz, Austria

 

Reviewers: The editorial board counts on reviewers from the ISMIR community, who are crucial to the success of the journal. To become a reviewer, please register here http://tismir.ubiquitypress.com/author/register/reviewer/

Journal Manager

Tim Wakeford, Ubiquity Press, United Kingdom

Contact

tismir@ismir.net

 

Website

http://tismir.ismir.net/

Leave a comment

Filed under publications, research, Uncategorized

Paper and dataset for Choir Singing Analysis, presented at ICMPC-ESCOM

Last week, Helena Cuesta, one of the PhD students I am working with, attended the 15th International Conference on Music Perception and Cognition and 10th triennial conference of the European Society for the Cognitive Sciences of Music in Graz (Austria). She presented the following paper in the poster session, as well as a contribution to the proceedings:

Cuesta, H., Gómez, E., Martorell, A., Loáiciga, F. Analysis of Intonation in Unison Choir Singing.

ICMPC/ESCOM is a very multidisciplinary conference, bringing together people from very different fields related to music such as music psychology, music perception, neuroscience, music theory, or music information retrieval.

The study investigates several expressive characteristics of unison choir singing, focusing on how singers blend together and interact with each other in terms of fundamental frequency dispersion, intonation, and vibrato. They also present an open dataset of choral singing that is available here, and was created in collaboration with the Anton Bruckner Choir (Barcelona).

This is a picture of the recording session. This work is being carried out in the context of two research projects: CASAS and TROMPA.

bruckner.png

 

Leave a comment

Filed under datasets, publications, research

Keynote speech @ 2018 Joint Workshop on Machine Learning for Music – ICML

raphael-andres-430356-unsplash-edit

I am honored to be keynote speaker at a joint workshop on Machine Learning for Music that will take place this summer in Stockholm. I spent part of my PhD in KTH in Stockholm and it became one of my favourite places, so it is always great to visit. T

In my keynote, I presented some current research of our lab on the analysis and synthesis of singing. In particular, we summarized some recent advances on a set of tasks related to the processing of singing using state-of-the-art deep learning techniques. We discussed their achievements in terms of accuracy and sound quality, and the current challenges, such as availability of data and computing resources. We also discussed the impact that these advances do and will have on listeners and singers when they are integrated in commercial applications.

There is paper related to this research which can be found in arxiv.

 

Leave a comment

Filed under events, keynotes

LifeSoundTrack: música, tecnología, recuerdos y alzheimer / music, technology, memories and alzheimer

English version below!

A veces los proyectos más enriquecedores son los que realizas con menos recursos.

En éste proyecto utilizamos las tecnologías de recuperación de la información musical para encontrar la banda sonora de la vida de personas mayores españolas. Ésto ha presentado varios retos de investigación y tecnológicos. En concreto hemos podido observar el sesgo de los sistemas actuales de recomendación musical tanto en que sólo se centran en un repertorio musical popular y sus interfaces y descriptores musicales están pensados para usuarios jóvenes. Esto aplica a:

  • Las taxonomías de géneros musicales.
  • Los idiomas predominantes en las canciones y su etiquetado.
  • Los descriptores y playlists que se utilizan (e.g. música de fiesta).
  • La dificultad para tener los derechos de dar una canción a éstas personas para que la escuchen.
  • La dificultad de tener dispositivos fáciles de usar y dirigidos a éste tipo de personas.

Por tanto casi hemos tenido que empezar desde cero! Éste es el repositorio github que hemos creado (abierto): https://github.com/MTG/lifesoundtrack

Y si alguien quiere obtener su banda sonora puede hacerlo en bandasonoravital.upf.edu 

Atención! El sistema está pensado para personas mayores que han nacido o viven en España, por lo que utilizadlo sobre todo si cumplís éstos requisitos o para alguna persona que conozcáis que los cumpla.

Ha sido un privilegio poder colaborar con la Fundación Pasqual Maragall, La Fundación AVAN y la Escuela La Salud de Sabadell en éste piloto con personas que padecen la enfermedad de Alzheimer. Sobre todo ha sido genial trabajar con Nina, Anna, Carolina, y los chicos de La Salut. Aquí podéis ver un vídeo que me encanta y resume muy bien el proyecto:

Para más información no os perdáis el documental Sense Ficció que emitirá TV3 el día 8 de Mayo por la noche!!!

Más información del proyecto en la web del MTG y de la Fundación Pasqual Maragall.

Y las bandas sonoras se pueden generar en http://bandasonoravital.upf.edu/ 

¡Gracias a todo el equipo! En especial al grupo del MTG: Perfe Herrera, Felipe Navarro, Olga Slizovskaia, por su tiempo en éste proyecto donde no hemos tenido financiación específica.

auriculares

Sometimes the most enriching projects are those that you do with less resources and funding.

In this project we use music information retrieval (MIR) technologies to find the life soundtrack of Spanish elderly people. Starting from a questionnaire where we ask about biographical information and one´s relationship with music, we build a playlist looking at several sources such as spotify or youtube. 

This has presented several research and technological challenges. Specifically, we have been able to observe the bias of current music recommendation systems, as they focus on a popular musical repertoire and their interfaces and musical descriptors are designed for young users. This applies to:

  • The taxonomies of musical genres.
  • The predominant languages ​​in the songs and their labeling.
  • The descriptors and playlists that are used (e.g., what does happy music mean).
  • The difficulty to have the rights to give a song to these people to be heard.
  • The difficulty of having easy-to-use devices aimed at this type of person.

Therefore we almost had to start from scratch! This is the github repository that we have created (open): https://github.com/MTG/lifesoundtrack 

And if someone wants to get their soundtrack you can do it at bandasonoravital.upf.edu
Attention! The system is designed for seniors who have been born or live in Spain, so use it especially if you meet these requirements or for someone you know who complies.

It has been a privilege to be able to collaborate with the Pasqual Maragall Foundation, the AVAN Foundation and the Health School of Sabadell in this pilot with people suffering from Alzheimer’s disease. Above all it has been great to work with Nina, Anna, Carolina, and the guys from La Salut. Here you can see a video that I love and sums up the project very well:

For more information, do not miss the Sense Ficció documentary that will broadcast TV3 on May 8 at night !!!

More information on the project on the MTG website and the Pasqual Maragall Foundation. Thanks to all the team! Especially to the MTG group: Perfe Herrera, Felipe Navarro, Olga Slizovskaia, for their time in this project where we have not had specific funding.

Leave a comment

Filed under outreach, research, software

Interview at ACM SIGMM Records and the need for interdisciplinary research

The March Issue (Vol.10, Issue 1) of ACM SIGMM Records (News for the Multimedia Community) is out and it includes an interview of myself for the interdisciplinary column, kindly chaired by Cynthia Liem and Jochen Huber.

It is awesome already that there is an interdisciplinary column at SIGMM, recognising the challenges and also the potential of interdisciplinary research and insights as a way to have a comprehensive understanding, in this case, of multimedia computing. I was very pleased to ask questions about my experience in the MIR field and about diversity and interdisciplinarity.

You can read the interview and other interesting content here.

Leave a comment

Filed under outreach, press, research, womenTIC

Presentación en Andalucía Digital Week 2018

El lunes tuve la oportunidad de presentar el proyecto HUMAINT con la charla “La influencia de los algoritmos en nuestras mentes y en nuestras decisiones” en la primera edición de Andalucía Digital Week , un evento que nace para visibilizar el sector de las TICs en Andalucía como punto de encuentro de empresas, administración y la sociedad. Entre las iniciativas más interesante está la jornada WOMANDIGITAL que visibiliza el papel de las mujeres en la transformación digital. No dejéis de firmar el manifiesto que hay en la web!!!

En mi intervención presenté el proyecto HUMAINT que lidero dentro del Centre for Advanced Studies, Joint Research Centre de la Comisión Europea, en el que estudiaremos el impacto que tendrá la inteligencia artificial en el comportamiento humano, principalmente en la toma de decisiones y en nuestras capacidades cognitivas. Fue muy interesante presentar por primera vez en Sevilla el proyecto y tener comentarios muy interesantes de los asistentes, con la pena de que justo al acabar me tuve que ir volando para el aeropuerto. Aquí hay algunas fotos que se pusieron en twitter!

 

 

Leave a comment

Filed under events, keynotes, outreach, womenTIC

Interview with Tara Chklovski from Iridescent Learning

Las December I had the chance to virtually meet Tara Chklovski, who is the CEO of Iridescent Learning,  a science education nonprofit organization in the United States.

logo-iridescent-300x133

Their mission, according to their website, is “to empower the world´s underrepresented young people, especially girls, through engineering and technology to become innovators and leader”.

They just launched (with AAAI, a leading organization in Artificial Intelligence) — the first, global Curiosity Machine AI Family Challenge – a two-stage competition for 20,000 underserved 3rd-8th grade students and parents (especially mothers) to use AI technologies and tools (sensors, data analysis tools) to solve problems in their communities (along the tracks of health, energy, food, transportation, education, public safety and civic engagement).

Of course, I thought: “what a wonderful mission and challenges!” How can I help? So she proposed me to do an interview where I could provide my thoughts on AI and its impact, targeting this particular audience.

 

It took me some weeks to complete it, as I have been very busy moving to another city, but I managed to answer the questions I hope in an interesting way. I hope this can also help to disseminate our research community, ISMIR, and its efforts to increase inclusion and variety, WiMIR, which I am very proud to contribute to.

I hope you like the interview, which can be found here.

Leave a comment

Filed under outreach, press, womenTIC

Hello world: el ¿primer? álbum ¿creado? por ¿inteligencia? artificial. 

Hoy en el programa de Radio Clásica, Longitud de Onda #LDOnda explicaré cuál ha sido el proceso de elaboración de “Hello world“, que según sus creadores es el primer álbum creado por inteligencia artificial.

Podréis escuchar el programa hoy en directo o más tarde en los podcasts de Longitud de Onda. También podéis escuchar otros programas en los que he participado aquí.

dv5rt33w0aevg3z

Leave a comment

Filed under press, teaching

“Contratados por un robot”: press article about artificial intelligence

Contratados por un robot“, an article published in el Periódico on November 11th, 2017 about artificial intelligence, mentions the new project HUMAINT that I will start leading in 2018.

Read the article in el Periodico.

Leave a comment

Filed under press

¿Qué impacto tendrá la inteligencia artificial en el comportamiento humano?

Noticia en la web de la UPF sobre el nuevo proyecto HUMAINT

Podéis leerla aquí.

Leave a comment

Filed under press, research

HUMAINT: New research project on Human Behaviour and Machine Intelligence @ Joint Research Centre (EC) – We are hiring!

Postdoc application deadline: OCTOBER 26, 2017

casmihb

I am excited to lead a novel research initiative inside the European Commission’s Joint Research Centre, on the topic of machine learning and human behaviour.

The Joint Research Centre (JRC) is the European Commission’s science and knowledge service which employs scientists to carry out research in order to provide independent scientific advice and support to EU policy. The JRC Centre for Advanced Studies (JRC-CAS) was established to enhance the JRC’s capabilities to meet emerging challenges at the science-policy interface. JRC-CAS is now launching a three year interdisciplinary project to understand the potential impact of machine learning in human behaviour and societal welfare. It will be carried out at JRC centre in Seville, Spain. There will be close collaboration with the Music Technology Group and the Department of Information and Communication Technologies of Universitat Pompeu Fabra  in Barcelona, Spain.

The HUMAINT project will (1) provide a scientific understanding of machine vs human intelligence; (2) analyze the influence of machine learning algorithms into human behaviour (3) investigate to what extent these findings should influence the European regulatory framework. Given my research expertise, music will be an important use case to address.

In the context of this project, three postdoc positions in the area of machine learning and human behaviour are open for appointment from January 1, 2018, at the Joint Research Centre (European Commission) in Seville, Spain. The fully funded positions are available for a period of three years. Particular areas of interests:

  • Fairness, accountability, transparency, explainability of machine learning methods.

  • Social, ethical and economic aspects of artificial intelligence.

  • Human-computer interaction and human-centered machine learning.

  • Digital and behavioural economy.

  • Application domains: music and arts, social networks, health, transport, energy.

We are looking for highly motivated, independent, and outstanding postdoc candidates with a strong background in machine learning and/or human behaviour. An excellent research track record, ability to communicate research results and involvement in community initiatives is expected. Candidates should have EU/EEA citizenship.

The JRC offers an enriching multi-cultural and multi-lingual work environment with lifelong learning and professional development opportunities, and close links to top research organisations and international bodies around the world. Postdoctoral researchers receive a competitive salary and excellent working conditions, and will define their own research agenda inline with the project goals.

JRC-Seville is located in Cartuja 93 scientific and technological park. Seville is the fourth-largest city in Spain. With more than 30 centuries of history (gateway of America for two centuries, main actor in the first circumnavigation of the Earth),  three UNESCO World Heritage Sites, and privileged climate, it combines its historical and touristic character with a consolidated economic development and innovation potential.

We are also open for collaborations with external researchers, as one of our goals is to build an expert network in the topics.

You may obtain further information about the scientific aspects of the positions from Dr. Emilia Gómez (project scientific leader: emilia.gomez@upf.edu, using the subject [humaint]) and at the following web pages http://recruitment.jrc.ec.europa.eu/?site=SVQ&type=AX&category=FGIV and https://ec.europa.eu/jrc/en/working-with-us/jobs.

YOU CAN APPLY HERE BY OCTOBER 26, 2017 (first application round).

By the way, the positions are in Seville, the city I was born, one of the most beautiful places in the world 🙂

Leave a comment

Filed under projects, research

New open-access journal Transactions of ISMIR, open for submissions

I am happy to announce that the International Society for Music Information Retrieval launched the Transactions of the International Society for Music Information Retrieval, the open access journal of the ISMIR society at Ubiquity press. I am serving as Editor-in-Chief, together with Simon Dixon and Anja Volk.

TISMIR publishes novel scientific research in the field of music information retrieval (MIR).

We welcome submissions from a wide range of disciplines: computer science, musicology, cognitive science,  library & information science and electrical engineering.

We currently accept submissions.

View our submission guidelines for more information.

TISMIR

Leave a comment

Filed under publications, research