2006:Poster Session and Discussion Topics

Introduction

This is a quick and dirty wiki page for MIREX folks to post their names and tentative poster titles for the Wednesday, 11 October 2006, MIREX 2006 lunch/poster session from 12:00-14:00 at ISMIR 2006.

Under the POSTER INFORMATION section, I would like you to simply add your name, your real or tentative poster title/topic and an indication that you will be able to attend the poster session. If you would like to have poster presence, but cannot be at the poster session, please contact me ASAP (jdownie@uiuc.edu) as we might be able to work something out.

Under the DISCUSSION TOPICS section, please post all your ideas, suggestions, gripes, comments that you think would help make the MIREX Results and Planning Plenary to be held Thursday, 12 October 2006 in the afternoon after lunch as productive and informative as possible.

POSTER INFORMATION

Downie, J. Stephen and the IMIRSEL Team, "MIREX 2006 Overview", Yes!

Pohle, Tim H., "Post Processing Music Similarity Queries", At least partly (talk immediately after the poster session).

Uitdenbogerd, Alexandra L., "Local Alignment Variations for Symbolic Music Retrieval", yes

Ferraro, Pascal and Hanna, Pierre, "Polyphonic Adaptation of Edit-Distance Algorithms", yes

Pampalk, E., "Audio-Based Music Similarity and Retrieval: Combining a Spectral Similarity Model with Information Extracted from Fluctuation Patterns", yes

Typke, Rainer/Wiering, Frans/Veltkamp, R. C.: "Symbolic Melodic Similarity with transportation distances", yes

Dixon, Simon, `"BeatRoot and Simple Spectral Onset Detection", yes

Lidy, Thomas, and Rauber, Andreas, "Computing Statistical Spectrum Descriptors for Audio Music Similarity and Retrieval", yes

Jang, Roger, "Simple But Effective Methods for QBSH at MIREX 2006", yes

I. Antonopoulos, A. Pikrakis and S. Theodoridis, "A tempo extraction algorithm for raw audio recordings", yes

Ellis, Dan, "Cover Song ID with Beat-Synchronous Chroma Features" (includes some overview and analysis of cover song task), no (sob!)

DISCUSSION TOPICS

J. Stephen Downie's Suggestions

1. Should we have some kind of "mid-year" planning meeting?
2. The never-ending shortage of test sets!
3. Starting novel tasks ASAP (see above)
4. Coding for robustness (the dirty data problem)
5. Task v. Contest issue
6. Rewarding/recognizing strong submissions (see above)

A. L. Uitdenbogerd's Suggestions

1. Standardisation of input/output requirements across years
2. Participation by data providers (training versus test/evaluation)

Simon Dixon`s Suggestion

I agree with standardisation across years, in particular that all entries should be run for at least two successive years to preserve continuity.

A. Pienimäki's Suggestions

1. Coordination and information sharing between the IMIRSEL team and task organisers (incl. the mid-year planning meeting)
2. Scheduling issues

Thomas Lidy's Suggestion

1. Other forms of (pre-task and task detail) discussion & communication than a Mailing List