Doctoral fellow CONICET @ Depto. De Computación, FCEN UBA / ICC, CONICET
Mail: lpepino at dc uba ar
Bio
Acoustics engineer, Universidad Nacional de Tres de Febrero (UNTREF)
Doctoral fellow under the supervision of Luciana Ferrer and Pablo Riera, and assistant teacher in the “Machine learning” course of the Master in Data Exploitation and Knowledge Discovery (UBA). His doctoral work consists in applying deep learning models to learn useful audio representations which can be used in low-resource scenarios.
Research interests:
- Acoustic signal processing
- Speech processing
- Deep learning
- Representation learning
- Transfer learning
Publications
Link to my Google Scholar
Selected articles:
- Leonardo Pepino, Pablo Riera, Luciana Ferrer, “Study of positional encoding approaches for Audio Spectrogram Transformers” arXiv preprint:2110.06999. Submitted to ICASSP (2022).
- Leonardo Pepino, Pablo Riera, Luciana Ferrer, “Emotion Recognition from Speech Using Wav2vec 2.0 Embeddings”. Interspeech (2021).
- Leonardo Pepino, Pablo Riera, Luciana Ferrer, Agustín Gravano, “Fusion approaches for emotion recognition from speech using acoustic and text-based features” ICASSP (2020)
- Leonardo Pepino, Pablo Riera, Lara Gauder, Agustín Gravano, Luciana Ferrer, “Detecting Distrust Towards the Skills of a Virtual Assistant Using Speech” arXiv:2007.15711 (2020)
Grants and awards
- 1st place in the first Cochlear Implant Hackathon (2021).
- Ing. Isidoro Marín award, given to outstanding recently graduated argentinian young engineers (2020).
- My doctoral project was awarded the Google Faculty Research Award 2019-2020 and the Amazon Research Awards 2019 in the field of Machine Perception.
- Doctoral fellowship for strategic topics CONICET 2019.