no code implementations • 14 Sep 2023 • Navin Raj Prabhu, Bunlong Lay, Simon Welker, Nale Lehmann-Willenbrock, Timo Gerkmann
Subsequently, at inference, a target emotion embedding is employed to convert the emotion of the input utterance to the given target emotion.
no code implementations • 2 Jun 2023 • Navin Raj Prabhu, Nale Lehmann-Willenbrock, Timo Gerkmann
In this work, we specifically focus on in-the-wild emotion conversion where parallel data does not exist, and the problem of disentangling lexical, speaker, and emotion information arises.
1 code implementation • 30 Sep 2022 • Navin Raj Prabhu, Nale Lehmann-Willenbrock, Timo Gerkman
Instead of a Gaussian, we model the annotation distribution using Student's t-distribution, which also accounts for the number of annotations available.
1 code implementation • 25 Jul 2022 • Navin Raj Prabhu, Nale Lehmann-Willenbrock, Timo Gerkmann
To address this, these emotion annotations are typically collected by multiple annotators and averaged across annotators in order to obtain labels for arousal and valence.
1 code implementation • 7 Oct 2021 • Navin Raj Prabhu, Guillaume Carbajal, Nale Lehmann-Willenbrock, Timo Gerkmann
At training, the network learns a distribution of weights to capture the inherent uncertainty related to subjective arousal annotations.