WP 2 -- Meta-Analysis of the MIR Discipline
From MIReS
Contents |
Objectives
The primary objective of this WP is to provide an in-depth, critical overview of past and current research trends in MIR. More particularly, a critical reflection will be conducted, as a collaborative effort, on the evolution of the field and on achievements and pitfalls from its beginning to today, both in terms of research findings but also employed methodologies, societal impacts and exchanges -inward and outward- with neighboring fields of science.
This meta-analysis of the field will be conducted in coordination with WP3 (MIR Roadmap) and will aim at providing insights on how the field should evolve towards sustainability. A second objective of this WP is to make this meta-analysis of MIR widely diffused, and suitable for different readerships, from specialized MIR researchers to beginners and students, and to different users of MIR technologies.
The involvement of the MIR community at large and relevant thinkers of the field in particular, will be instrumental towards this objective.
Description of work and role of partners
In this WP, the consortium will write an initial review on MIR research trends, with special focus on benchmarking activities. This review will serve as “bootstrap” for further edition by the consortium, together with the involvement of the MIR community at large.
Task 2.1: Organise the edition process of a web-based, collaborative, state-of-the-art documentation of MIR research
In this WP, we will produce a structured collection of documents that will encompass a complete overview of MIR. This documentation will have to be exhaustive, to keep updated with the rapid evolution of the field, and to have a significant impact on the MIR community (i.e. serve effectively as reference hub for MIR documentation). For this to be possible, we will set up a web-based repository on the MIReS web site (see WP4), very much inspired by the structure of existing repositories such as Wikipedia, including different formats for documentations, which will include concise articles, lecture slides, discussion forums, thematic news feeds, links to relevant online literature and multimedia material, etc. By the end of the project, and partly during its course, this repository will be editable by the community itself.
The role of the consortium will be to (1) define the main structure and document formats of this repository (in particular, QMUL will contribute here as a continuation of web presence efforts in WP4), (2) include initial contributions (see Task 2.2), (3) reach out to the MIR community and in particular key researchers beyond the consortium for their involvement (in particular, QMUL and UPF-MTG will contribute here as continuation of work in WPs 4 and 5), (4) define a clear set of rules and responsibilities for editing/monitoring/reviewing/mediation permissions by members of the community, and (5) ensure that any information released is available on open networks and covered by Creative Commons licenses. This will be achieved as a collaborative process including all partners of the consortium.
Partners involved: INESC PORTO overall coordination and execution of tasks described above, all partners report and contribute.
Task 2.2: Review of MIR research trends
The initial contributions to the web-based repository will be produced by the consortium itself, it will provide the required input to WP3, and will serve as "bootstrap" for further revisions by the community at large. The writing process will encompass reviewing the following three aspects, and writing responsibilities will be distributed among consortium partners: MIR research topics: audio content analysis, context data analysis, user profiling, musical representations, research tools, etc. (IRCAM and OSGK-OFAI expertise on those topics will be of particular relevance)
MIR applications: search, recommendation, digital libraries, rights management, music visualisation, music promotion, music networking, music education/training, interactive arts, etc. (IRCAM and BMAT expertise will be instrumental here)
MIR challenges: multiculturalism, semantic gap, support to creativity, multimodality, multidisciplinarity, etc. (here UPF-MTG expertise will be instrumental)
Involvement of key MIR researchers/experts in this task will be facilitated by the organization of meetings with members of the consortium, for instance during events organized in WPs 4 and 5, but also possibly as short-term individual actions.
Task 2.3: Assess benchmarking efforts in MIR
This task is conceived as a series of efforts to aid the writing of MIReS standards in Work Package 3, Task 3.2, and documentation part B1.1.2.1. Advances in particular fields of sciences are often determined by how well scientists can evaluate their progress. For this reason, we will dedicate particular attention to a critical assessment of evaluation efforts in MIR, including MIREX (http://www.music-ir.org/mirex/), MUSICLEF (http://www.promise-noe.eu/web/guest), and Audio Description Contest (http://ismir2004.ismir.net/ISMIR_Contest.html).
We will also monitor other evaluation and benchmarking efforts from other, more established fields of science, such as Information Retrieval, Natural Language Processing, etc.: TREC (http://trec.nist.gov/), CLEF (http://www.clef-campaign.org/), and TRECVID (http://trecvid.nist.gov/).
Partners involved: INESC PORTO overall coordination and execution of tasks described above, all partners report and contribute.
Deliverables
Deliverable 2.1: Documentation hub
Database for the web-based documentation hub [month 6]
Deliverable 2.2: Written report
Written report on initial review of MIR research trends, MIR benchmarking, and definition of editing structure of the web-based documentation hub, to be used by WP3 [month 9]
Deliverable 2.3: Final version of the web-based documentation hub
Final version of the web-based documentation hub [month 18]
