Acoustic Scene Classification with Mismatched Recording Devices

Paul Primus; David Eitelsebner

Acoustic Scene Classification with Mismatched Recording Devices

Paul Primus, David Eitelsebner

Institute of Computational Perception

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings

Abstract

This technical report describes CP-JKU Student team’s approach for Task 1 - Subtask B of the DCASE 2019 challenge. In this context, we propose two loss functions for domain adaptation to learn invariant representations given time-aligned recordings. We show that these methods improve the classification performance on our cross-validation, as well as performance on the Kaggle leader board, up to three percentage points compared to our baseline model. Our best scoring submission is an ensemble of eight classifiers.

Original language	English
Title of host publication	DCASE 2019 Technical Report
Number of pages	4
Publication status	Published - 2020

Fields of science

202002 Audiovisual media
102 Computer Sciences
102001 Artificial intelligence
102003 Image processing
102015 Information systems

JKU Focus areas

Digital Transformation

Cite this

@inproceedings{a3176bf784eb4f1a911a8232b4343387,

title = "Acoustic Scene Classification with Mismatched Recording Devices",

abstract = "This technical report describes CP-JKU Student team{\textquoteright}s approach for Task 1 - Subtask B of the DCASE 2019 challenge. In this context, we propose two loss functions for domain adaptation to learn invariant representations given time-aligned recordings. We show that these methods improve the classification performance on our cross-validation, as well as performance on the Kaggle leader board, up to three percentage points compared to our baseline model. Our best scoring submission is an ensemble of eight classifiers.",

author = "Paul Primus and David Eitelsebner",

year = "2020",

language = "English",

booktitle = "DCASE 2019 Technical Report",

}

TY - GEN

T1 - Acoustic Scene Classification with Mismatched Recording Devices

AU - Primus, Paul

AU - Eitelsebner, David

PY - 2020

Y1 - 2020

N2 - This technical report describes CP-JKU Student team’s approach for Task 1 - Subtask B of the DCASE 2019 challenge. In this context, we propose two loss functions for domain adaptation to learn invariant representations given time-aligned recordings. We show that these methods improve the classification performance on our cross-validation, as well as performance on the Kaggle leader board, up to three percentage points compared to our baseline model. Our best scoring submission is an ensemble of eight classifiers.

AB - This technical report describes CP-JKU Student team’s approach for Task 1 - Subtask B of the DCASE 2019 challenge. In this context, we propose two loss functions for domain adaptation to learn invariant representations given time-aligned recordings. We show that these methods improve the classification performance on our cross-validation, as well as performance on the Kaggle leader board, up to three percentage points compared to our baseline model. Our best scoring submission is an ensemble of eight classifiers.

M3 - Conference proceedings

BT - DCASE 2019 Technical Report

ER -