3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement
Artificial neural networks (ANNs) have become an important part of signal processing research. While ANNs outperform model-based signal processing methods in many applications, their internal processing often remains unclear. In this thesis, a framework for analyzing the signal processing performed...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-03-01
|
| Series: | Science Talks |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S277256932500012X |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823856695619616768 |
|---|---|
| author | Annika Briegleb |
| author_facet | Annika Briegleb |
| author_sort | Annika Briegleb |
| collection | DOAJ |
| description | Artificial neural networks (ANNs) have become an important part of signal processing research. While ANNs outperform model-based signal processing methods in many applications, their internal processing often remains unclear. In this thesis, a framework for analyzing the signal processing performed by ANN-based filters for multichannel speech enhancement is proposed. By designing specific training and test scenarios that allow to associate each time frame with certain information, e.g., spatial cues, and using low-cost analysis tools such as clustering, interpretable information can be extracted from the hidden features of the ANN. The proposed framework allows to assess whether and where spatial information is represented inside the ANN, answering the question whether these ANNs exploit spatial cues in addition to spectral information. Furthermore, the impact of the choice of training target on the functionality and interpretability of the ANN is considered. By applying the proposed analysis tools to two conceptually different speech enhancement frameworks, it is shown that the amount of spatial information extracted inside the ANN varies depending on the training target and the test scenario. The insights from this thesis help to assess the signal processing capabilities of ANNs and allow to make informed decisions when configuring, training, and deploying ANNs. |
| format | Article |
| id | doaj-art-c44d4789c90146bf8d761cd9e6846873 |
| institution | Kabale University |
| issn | 2772-5693 |
| language | English |
| publishDate | 2025-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Science Talks |
| spelling | doaj-art-c44d4789c90146bf8d761cd9e68468732025-02-12T05:33:08ZengElsevierScience Talks2772-56932025-03-01131004303MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancementAnnika Briegleb0Multimedia Communications and Signal Processing, Friedrich-Alexander-Universität Erlangen-Nürnberg, Cauerstr. 7, 91058 Erlangen, GermanyArtificial neural networks (ANNs) have become an important part of signal processing research. While ANNs outperform model-based signal processing methods in many applications, their internal processing often remains unclear. In this thesis, a framework for analyzing the signal processing performed by ANN-based filters for multichannel speech enhancement is proposed. By designing specific training and test scenarios that allow to associate each time frame with certain information, e.g., spatial cues, and using low-cost analysis tools such as clustering, interpretable information can be extracted from the hidden features of the ANN. The proposed framework allows to assess whether and where spatial information is represented inside the ANN, answering the question whether these ANNs exploit spatial cues in addition to spectral information. Furthermore, the impact of the choice of training target on the functionality and interpretability of the ANN is considered. By applying the proposed analysis tools to two conceptually different speech enhancement frameworks, it is shown that the amount of spatial information extracted inside the ANN varies depending on the training target and the test scenario. The insights from this thesis help to assess the signal processing capabilities of ANNs and allow to make informed decisions when configuring, training, and deploying ANNs.http://www.sciencedirect.com/science/article/pii/S277256932500012XExplainable AIDNN interpretabilityMultichannel speech enhancementSpatial filtering |
| spellingShingle | Annika Briegleb 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement Science Talks Explainable AI DNN interpretability Multichannel speech enhancement Spatial filtering |
| title | 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement |
| title_full | 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement |
| title_fullStr | 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement |
| title_full_unstemmed | 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement |
| title_short | 3MT Competition (EUSIPCO2024): A peek into the black box: Insights into the functionality of complex-valued neural networks for multichannel speech enhancement |
| title_sort | 3mt competition eusipco2024 a peek into the black box insights into the functionality of complex valued neural networks for multichannel speech enhancement |
| topic | Explainable AI DNN interpretability Multichannel speech enhancement Spatial filtering |
| url | http://www.sciencedirect.com/science/article/pii/S277256932500012X |
| work_keys_str_mv | AT annikabriegleb 3mtcompetitioneusipco2024apeekintotheblackboxinsightsintothefunctionalityofcomplexvaluedneuralnetworksformultichannelspeechenhancement |