Felix Weninger
Research Subjects
Felix Weninger has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Björn Schuller
Technische Universität München, Munich, Germany
5 shared publications Anton Batliner
Mustererkennung Labor, Friedrich-Alexander-Universität Erlangen-Nürnberg
1 shared publicationDino Seppi
ESAT, Katholieke Universiteit Leuven
1 shared publication- Emmanuel Vincent
Université de Lorraine CNRS F-54000 Nancy Inria LORIA France
1 shared publication - Florian Eyben
Technische Universität München
1 shared publication Hakan Erdogan
Mitsubishi Electric Research Laboratories
1 shared publication
Latest Publications
- Top Cited Papers
- 15 August 2015
- Published by Springer Nature
- p. 91-99
Speech Enhancement with LSTM Recurrent Neural Networks and its Application to Noise-Robust ASR
- 15 August 2015
- Published by Springer Nature
- p. 91-99
Top Cited PapersSpeech Enhancement with LSTM Recurrent Neural Networks and its Application to Noise-Robust ASR
- Open Access
Deep Unfolding: Model-Based Inspiration of Novel Deep Architectures
Model-based methods and deep neural networks have both been tremendously successful paradigms in machine learning. In model-based methods, problem domain knowledge can be built into the constraints of the model, typically at the expense of difficulties during inference. In contrast, deterministic deep neural networks are constructed in such a way that inference is straightforward, but their architectures are generic and it is unclear how to incorporate knowledge. This work aims to obtain the advantages of both approaches. To do so, we start with a model-based approach and an associated inference algorithm, and \emph{unfold} the inference iterations as layers in a deep network. Rather than optimizing the original model, we \emph{untie} the model parameters across layers, in order to create a more powerful network. The resulting architecture can be trained discriminatively to perform accurate inference within a fixed network size. We show how this framework allows us to interpret conventional networks as mean-field inference in Markov random fields, and to obtain new architectures by instead using belief propagation as the inference algorithm. We then show its application to a non-negative matrix factorization model that incorporates the problem-domain knowledge that sound sources are additive. Deep unfolding of this model yields a new kind of non-negative deep neural network, that can be trained using a multiplicative backpropagation-style update algorithm. We present speech enhancement experiments showing that our approach is competitive with conventional neural networks despite using far fewer parameters.
Open AccessDeep Unfolding: Model-Based Inspiration of Novel Deep Architectures
Model-based methods and deep neural networks have both been tremendously successful paradigms in machine learning. In model-based methods, problem domain knowledge can be built into the constraints of the model, typically at the expense of difficulties during inference. In contrast, deterministic deep neural networks are constructed in such a way that inference is straightforward, but their architectures are generic and it is unclear how to incorporate knowledge. This work aims to obtain the advantages of both approaches. To do so, we start with a model-based approach and an associated inference algorithm, and \emph{unfold} the inference iterations as layers in a deep network. Rather than optimizing the original model, we \emph{untie} the model parameters across layers, in order to create a more powerful network. The resulting architecture can be trained discriminatively to perform accurate inference within a fixed network size. We show how this framework allows us to interpret conventional networks as mean-field inference in Markov random fields, and to obtain new architectures by instead using belief propagation as the inference algorithm. We then show its application to a non-negative matrix factorization model that incorporates the problem-domain knowledge that sound sources are additive. Deep unfolding of this model yields a new kind of non-negative deep neural network, that can be trained using a multiplicative backpropagation-style update algorithm. We present speech enhancement experiments showing that our approach is competitive with conventional neural networks despite using far fewer parameters.
- 1 April 2013
- Published by IGI Global Scientific Publishing in International Journal of Distance Education Technologies
- Vol. 11 (2) , 110-123
Words that Fascinate the Listener
In a large scale study on 843 transcripts of Technology, Entertainment and Design (TED) talks, the authors address the relation between word usage and categorical affective ratings of lectures by a large group of internet users. Users rated the lectures by assigning one or more predefined tags which relate to the affective state evoked in the audience (e. g., ‘fascinating’, ‘funny’, ‘courageous’, ‘unconvincing’ or ‘long-winded’). By automatic classification experiments, they demonstrate the usefulness of linguistic features for predicting these subjective ratings. Extensive test runs are conducted to assess the influence of the classifier and feature selection, and individual linguistic features are evaluated with respect to their discriminative power. In the result, classification whether the frequency of a given tag is higher than on average can be performed most robustly for tags associated with positive valence, reaching up to 80.7% accuracy on unseen test data.
- 1 April 2013
- Published by IGI Global Scientific Publishing in International Journal of Distance Education Technologies
- Vol. 11 (2) , 110-123
Words that Fascinate the Listener
In a large scale study on 843 transcripts of Technology, Entertainment and Design (TED) talks, the authors address the relation between word usage and categorical affective ratings of lectures by a large group of internet users. Users rated the lectures by assigning one or more predefined tags which relate to the affective state evoked in the audience (e. g., ‘fascinating’, ‘funny’, ‘courageous’, ‘unconvincing’ or ‘long-winded’). By automatic classification experiments, they demonstrate the usefulness of linguistic features for predicting these subjective ratings. Extensive test runs are conducted to assess the influence of the classifier and feature selection, and individual linguistic features are evaluated with respect to their discriminative power. In the result, classification whether the frequency of a given tag is higher than on average can be performed most robustly for tags associated with positive valence, reaching up to 80.7% accuracy on unseen test data.
- Top Cited PapersOpen Access
- 1 January 2013
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 4, 292
On the Acoustics of Emotion in Audio: What Speech, Music, and Sound have in Common
Without doubt, there is emotional information in almost any kind of sound received by humans every day: be it the affective state of a person transmitted by means of speech; the emotion intended by a composer while writing a musical piece, or conveyed by a musician while performing it; or the affective state connected to an acoustic event occurring in the environment, in the soundtrack of a movie, or in a radio play. In the field of affective computing, there is currently some loosely connected research concerning either of these phenomena, but a holistic computational model of affect in sound is still lacking. In turn, for tomorrow’s pervasive technical systems, including affective companions and robots, it is expected to be highly beneficial to understand the affective dimensions of “the sound that something makes,” in order to evaluate the system’s auditory environment and its own audio output. This article aims at a first step toward a holistic computational model: starting from standard acoustic feature extraction schemes in the domains of speech, music, and sound analysis, we interpret the worth of individual features across these three domains, considering four audio databases with observer annotations in the arousal and valence dimensions. In the results, we find that by selection of appropriate descriptors, cross-domain arousal, and valence regression is feasible achieving significant correlations with the observer annotations of up to 0.78 for arousal (training on sound and testing on enacted speech) and 0.60 for valence (training on enacted speech and testing on music). The high degree of cross-domain consistency in encoding the two main dimensions of affect may be attributable to the co-evolution of speech and music from multimodal affect bursts, including the integration of nature sounds for expressive effects.
- 1 January 2013
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 4, 292
Top Cited PapersOpen AccessOn the Acoustics of Emotion in Audio: What Speech, Music, and Sound have in Common
Without doubt, there is emotional information in almost any kind of sound received by humans every day: be it the affective state of a person transmitted by means of speech; the emotion intended by a composer while writing a musical piece, or conveyed by a musician while performing it; or the affective state connected to an acoustic event occurring in the environment, in the soundtrack of a movie, or in a radio play. In the field of affective computing, there is currently some loosely connected research concerning either of these phenomena, but a holistic computational model of affect in sound is still lacking. In turn, for tomorrow’s pervasive technical systems, including affective companions and robots, it is expected to be highly beneficial to understand the affective dimensions of “the sound that something makes,” in order to evaluate the system’s auditory environment and its own audio output. This article aims at a first step toward a holistic computational model: starting from standard acoustic feature extraction schemes in the domains of speech, music, and sound analysis, we interpret the worth of individual features across these three domains, considering four audio databases with observer annotations in the arousal and valence dimensions. In the results, we find that by selection of appropriate descriptors, cross-domain arousal, and valence regression is feasible achieving significant correlations with the observer annotations of up to 0.78 for arousal (training on sound and testing on enacted speech) and 0.60 for valence (training on enacted speech and testing on music). The high degree of cross-domain consistency in encoding the two main dimensions of affect may be attributable to the co-evolution of speech and music from multimodal affect bursts, including the integration of nature sounds for expressive effects.
- 27 April 2012
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 69 (3) , 267-277
Optimization and Parallelization of Monaural Source Separation Algorithms in the openBliSSART Toolkit
- 27 April 2012
- Published by Springer Nature in Journal of Signal Processing Systems
- Vol. 69 (3) , 267-277
Optimization and Parallelization of Monaural Source Separation Algorithms in the openBliSSART Toolkit
- Open Access
- 23 January 2011
- Published by Springer Nature in EURASIP Journal on Advances in Signal Processing
- Vol. 2011 (1) , 838790
Recognition of Nonprototypical Emotions in Reverberated and Noisy Speech by Nonnegative Matrix Factorization
- 23 January 2011
- Published by Springer Nature in EURASIP Journal on Advances in Signal Processing
- Vol. 2011 (1) , 838790
Open AccessRecognition of Nonprototypical Emotions in Reverberated and Noisy Speech by Nonnegative Matrix Factorization