… In this work we show that we can train a single learnable frontend that … music, audio events
and animal sounds, providing a general-purpose learned frontend for audio classification. To …

… In experiments on six audio classification tasks, our frontend matches the accuracy of LEAF
at 3… The quest for learnable audio frontends is not solved. Index Terms—audio classification, …

… ”LEAF: A LEarnable Audio Frontend for Audio Classification”, which serves as the foundation
for this thesis. When assessed across a range of tasks in a variety of audio domains, their …

… learnable audio frontends: SincNet [12], and LEAF [16], both in a supervised multi-label
classification … On a linear probe transfer task, learnable frontends significantly outperform fixed …

… front-ends. Second, we provide a comparison of feature selections for spoofing countermeasures,
with a special focus on learnable audio front-ends… the classification of audio front-ends …

… front-end and joint training of the front-end and classification … LFBE directly from 7 channel
audio by using the super… from beamformed audio, we pre-train the front-end layer with …

… These front-ends perform better than using traditional mel-filterbanks on various audio
classification tasks. Unlike existing work on learnable front-ends, SimPFs are nonparametric and …

… classification system with learnable front-end - UALF. Parameterized Gabor filters are used
to organize the learnable front-end … Audio preprocessing mainly includes frame splitting and …

… Audio Feature Extraction The audio features are extracted using the learnable front-end
for audio classification (LEAF) [19] model, which learns the features using components like …

… for eight audio classification tasks. We constructed a learnable front-end that can compute
the … values of adjacent elements is more important than the phase itself in audio classification. …