Pizarro Matías

PhD student, Machine Learning, Ruhr-Universtät Bochum

Joined

September 2023

Names

Pizarro Matías (Preferred)

Matias Patricio Pizarro Bustamante

Matias Pizarro

Matías P. Pizarro B.

Emails

****@gmail.com (Confirmed)

****@ruhr-uni-bochum.de (Confirmed)

Personal Links

Google Scholar

DBLP

ORCID

Career & Education History

PhD student

Machine Learning, Ruhr-Universtät Bochum (rub.de)

2020 – Present

Advisors, Relations & Conflicts

PhD Advisor

Asja Fischer

2020 – Present

PhD Advisor

Dorothea Kolossa

2020 – Present

Expertise

Automatic speech recognition

Present

Deepfake detection

Present

Uncertainty Quantification

Present

Publications

Reassessing Noise Augmentation Methods in the Context of Adversarial Speech
Pizarro Matías, Fischer Asja, Pizzi Karla
- 4th Symposium on Security and Privacy in Speech Communication
- Readers: Everyone
Comparative study on noise-augmented training and its effect on adversarial robustness in ASR systems
Fischer Asja, Pizarro Matías, Pizzi Karla
- Computer speech and language
- Readers: Everyone
Robustifying automatic speech recognition by extracting slowly varying features
Pizarro Matías, Kolossa Dorothea, Fischer Asja
- 2021 ISCA Symposium on Security and Privacy in Speech Communication
- Readers: Everyone
DistriBlock: Identifying adversarial audio samples by leveraging characteristics of the output distribution
Matias Patricio Pizarro Bustamante, Dorothea Kolossa, Asja Fischer
- UAI 2024 poster
- Readers: Everyone
Single-Model Attribution for Spoofed Speech via Vocoder Fingerprints in an Open-World Setting
Matias Pizarro, Mike Laszkiewicz, Dorothea Kolossa, Asja Fischer
- CoRR 2024
- Readers: Everyone
Leveraging characteristics of the output distribution for identifying adversarial audio examples
Matias Patricio Pizarro Bustamante, Dorothea Kolossa, Asja Fischer
- Submitted to ICLR 2024
- Readers: Everyone

Pizarro Matías

Names

Emails

Personal Links

Career & Education History

Advisors, Relations & Conflicts

Expertise

Publications

Reassessing Noise Augmentation Methods in the Context of Adversarial Speech

Comparative study on noise-augmented training and its effect on adversarial robustness in ASR systems

Robustifying automatic speech recognition by extracting slowly varying features

DistriBlock: Identifying adversarial audio samples by leveraging characteristics of the output distribution

Single-Model Attribution for Spoofed Speech via Vocoder Fingerprints in an Open-World Setting

Leveraging characteristics of the output distribution for identifying adversarial audio examples

Co-Authors