Learning, Probability and Logic: Toward a Unified Approach for Content-Based Music Information Retrieval. Crayencour, H. & Cella, C. Frontiers in Digital Humanities, 6(April):1–25, 2019. doi abstract bibtex Within the last fifteen years, the field of Music Information Retrieval (MIR) has made tremendous progress in the development of algorithms for organizing and analyzing the ever-increasing large and varied amount of music and music-related data available digitally. However, the development of content-based methods to enable or improve multimedia retrieval still remains a central challenge. In this perspective paper, we critically look at the problem of automatic chord estimation from audio recordings as a case study of content-based algorithms, and point out several bottlenecks in current approaches: expressiveness and flexibility are obtained to the expense of robustness and vice-versa; available multimodal sources of information are little exploited; modeling multi-faceted and strongly interrelated musical information is limited with current architectures; models are typically restricted to short-term analysis that does not account for the hierarchical temporal structure of musical signals. Dealing with music data requires the ability to handle both uncertainty and complex relational structure at multiple levels of representation. Traditional approaches have generally treated these two aspects separately, probability and learning being the standard way to represent uncertainty in knowledge, while logical representation being the standard way to represent knowledge and complex relational information. We advocate that the identified hurdles of current approaches could be overcome by recent developments in the area of Statistical Relational Artificial Intelligence (StarAI) that unifies probability, logic and (deep) learning. We show that existing approaches used in MIR find powerful extensions and unifications in StarAI, and we explain why we think it is time to consider the new perspectives offered by this promising research field.
@Article{ crayencour.ea2019-learning,
author = {Crayencour, Helene-Camille and Cella, Carmine-Emanuele},
year = {2019},
title = {Learning, Probability and Logic: Toward a Unified
Approach for Content-Based Music Information Retrieval},
abstract = {Within the last fifteen years, the field of Music
Information Retrieval (MIR) has made tremendous progress
in the development of algorithms for organizing and
analyzing the ever-increasing large and varied amount of
music and music-related data available digitally. However,
the development of content-based methods to enable or
improve multimedia retrieval still remains a central
challenge. In this perspective paper, we critically look
at the problem of automatic chord estimation from audio
recordings as a case study of content-based algorithms,
and point out several bottlenecks in current approaches:
expressiveness and flexibility are obtained to the expense
of robustness and vice-versa; available multimodal sources
of information are little exploited; modeling
multi-faceted and strongly interrelated musical
information is limited with current architectures; models
are typically restricted to short-term analysis that does
not account for the hierarchical temporal structure of
musical signals. Dealing with music data requires the
ability to handle both uncertainty and complex relational
structure at multiple levels of representation.
Traditional approaches have generally treated these two
aspects separately, probability and learning being the
standard way to represent uncertainty in knowledge, while
logical representation being the standard way to represent
knowledge and complex relational information. We advocate
that the identified hurdles of current approaches could be
overcome by recent developments in the area of Statistical
Relational Artificial Intelligence (StarAI) that unifies
probability, logic and (deep) learning. We show that
existing approaches used in MIR find powerful extensions
and unifications in StarAI, and we explain why we think it
is time to consider the new perspectives offered by this
promising research field.},
doi = {10.3389/fdigh.2019.00006},
issn = {2297-2668},
journal = {Frontiers in Digital Humanities},
keywords = {audio,chord recognition,content-based,mir,music
information retrieval,music information retrieval
(MIR),statistical relational artificial,statistical
relational artificial intelligence},
mendeley-tags= {music information retrieval},
number = {April},
pages = {1--25},
volume = {6}
}
Downloads: 0
{"_id":"SLW6oWkcg5DHtqQHN","bibbaseid":"crayencour-cella-learningprobabilityandlogictowardaunifiedapproachforcontentbasedmusicinformationretrieval-2019","authorIDs":[],"author_short":["Crayencour, H.","Cella, C."],"bibdata":{"bibtype":"article","type":"article","author":[{"propositions":[],"lastnames":["Crayencour"],"firstnames":["Helene-Camille"],"suffixes":[]},{"propositions":[],"lastnames":["Cella"],"firstnames":["Carmine-Emanuele"],"suffixes":[]}],"year":"2019","title":"Learning, Probability and Logic: Toward a Unified Approach for Content-Based Music Information Retrieval","abstract":"Within the last fifteen years, the field of Music Information Retrieval (MIR) has made tremendous progress in the development of algorithms for organizing and analyzing the ever-increasing large and varied amount of music and music-related data available digitally. However, the development of content-based methods to enable or improve multimedia retrieval still remains a central challenge. In this perspective paper, we critically look at the problem of automatic chord estimation from audio recordings as a case study of content-based algorithms, and point out several bottlenecks in current approaches: expressiveness and flexibility are obtained to the expense of robustness and vice-versa; available multimodal sources of information are little exploited; modeling multi-faceted and strongly interrelated musical information is limited with current architectures; models are typically restricted to short-term analysis that does not account for the hierarchical temporal structure of musical signals. Dealing with music data requires the ability to handle both uncertainty and complex relational structure at multiple levels of representation. Traditional approaches have generally treated these two aspects separately, probability and learning being the standard way to represent uncertainty in knowledge, while logical representation being the standard way to represent knowledge and complex relational information. We advocate that the identified hurdles of current approaches could be overcome by recent developments in the area of Statistical Relational Artificial Intelligence (StarAI) that unifies probability, logic and (deep) learning. We show that existing approaches used in MIR find powerful extensions and unifications in StarAI, and we explain why we think it is time to consider the new perspectives offered by this promising research field.","doi":"10.3389/fdigh.2019.00006","issn":"2297-2668","journal":"Frontiers in Digital Humanities","keywords":"audio,chord recognition,content-based,mir,music information retrieval,music information retrieval (MIR),statistical relational artificial,statistical relational artificial intelligence","mendeley-tags":"music information retrieval","number":"April","pages":"1–25","volume":"6","bibtex":"@Article{ crayencour.ea2019-learning,\n author = {Crayencour, Helene-Camille and Cella, Carmine-Emanuele},\n year = {2019},\n title = {Learning, Probability and Logic: Toward a Unified\n Approach for Content-Based Music Information Retrieval},\n abstract = {Within the last fifteen years, the field of Music\n Information Retrieval (MIR) has made tremendous progress\n in the development of algorithms for organizing and\n analyzing the ever-increasing large and varied amount of\n music and music-related data available digitally. However,\n the development of content-based methods to enable or\n improve multimedia retrieval still remains a central\n challenge. In this perspective paper, we critically look\n at the problem of automatic chord estimation from audio\n recordings as a case study of content-based algorithms,\n and point out several bottlenecks in current approaches:\n expressiveness and flexibility are obtained to the expense\n of robustness and vice-versa; available multimodal sources\n of information are little exploited; modeling\n multi-faceted and strongly interrelated musical\n information is limited with current architectures; models\n are typically restricted to short-term analysis that does\n not account for the hierarchical temporal structure of\n musical signals. Dealing with music data requires the\n ability to handle both uncertainty and complex relational\n structure at multiple levels of representation.\n Traditional approaches have generally treated these two\n aspects separately, probability and learning being the\n standard way to represent uncertainty in knowledge, while\n logical representation being the standard way to represent\n knowledge and complex relational information. We advocate\n that the identified hurdles of current approaches could be\n overcome by recent developments in the area of Statistical\n Relational Artificial Intelligence (StarAI) that unifies\n probability, logic and (deep) learning. We show that\n existing approaches used in MIR find powerful extensions\n and unifications in StarAI, and we explain why we think it\n is time to consider the new perspectives offered by this\n promising research field.},\n doi = {10.3389/fdigh.2019.00006},\n issn = {2297-2668},\n journal = {Frontiers in Digital Humanities},\n keywords = {audio,chord recognition,content-based,mir,music\n information retrieval,music information retrieval\n (MIR),statistical relational artificial,statistical\n relational artificial intelligence},\n mendeley-tags= {music information retrieval},\n number = {April},\n pages = {1--25},\n volume = {6}\n}\n\n","author_short":["Crayencour, H.","Cella, C."],"key":"crayencour.ea2019-learning","id":"crayencour.ea2019-learning","bibbaseid":"crayencour-cella-learningprobabilityandlogictowardaunifiedapproachforcontentbasedmusicinformationretrieval-2019","role":"author","urls":{},"keyword":["audio","chord recognition","content-based","mir","music information retrieval","music information retrieval (MIR)","statistical relational artificial","statistical relational artificial intelligence"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://hmb.sampaio.me/bibliografia.bib.txt","creationDate":"2019-11-04T12:45:42.768Z","downloads":0,"keywords":["audio","chord recognition","content-based","mir","music information retrieval","music information retrieval (mir)","statistical relational artificial","statistical relational artificial intelligence"],"search_terms":["learning","probability","logic","toward","unified","approach","content","based","music","information","retrieval","crayencour","cella"],"title":"Learning, Probability and Logic: Toward a Unified Approach for Content-Based Music Information Retrieval","year":2019,"dataSources":["n6MFY2CscQLDpJ7nT","RFLDZw5KyJdadDXDm"]}