Carlos Manuel Ferreira Carvalho
From HLT@INESC-ID
Carlos Carvalho's academic journey in computer science commenced at Instituto Superior Técnico, Lisbon, where he achieved his Bachelor's and Master's degrees, specializing in Artificial Intelligence. His Master's thesis pioneered state-of-the-art end-to-end automatic speech recognition for European Portuguese, advancing beyond traditional HMM-based systems. Currently, he is pursuing a PhD, deepening his research in deep learning, speech recognition, with a keen focus on deep unsupervised representation learning (self-supervised learning), domain generalization, and addressing challenges in low-resource settings.
Carlos' work is focused on deep unsupervised learning, particularly self-supervised learning, for automatic speech recognition. His work centers on representation learning, aiming to improve how systems understand and process speech. He is also interested in the application of attention and memory mechanisms in deep learning, exploring how these can enhance speech recognition systems. His research contributes to advancements in machine learning and natural language processing.
Total Publications: 7
In Proceedings (Author)
- CAMÕES: A Comprehensive Automatic Speech Recognition Benchmark for European Portuguese
Carlos Manuel Ferreira Carvalho, Francisco Teixeira, Catarina Botelho, Anna Maria Pompili, Rubén Solera Ureña, Sérgio Paulo, Mariana Julião, Thomas Rolland, John Mendonça, Diogo Pereira, Isabel Trancoso, Alberto Abad,
2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU), IEEE., IEEE. © 20XX IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., December 2025 - Exploring Linear Variant Transformers and k-NN Memory Inference for Long-Form ASR
Carlos Manuel Ferreira Carvalho, Alberto Abad,
Proc. Interspeech, August 2025 - CS-FLEURS: A Massively Multilingual and Code-Switched Speech Dataset
Carlos Manuel Ferreira Carvalho,
Proc. Interspeech, August 2025 - AC-Mix: Self-Supervised Adaptation for Low-Resource Automatic Speech Recognition using Agnostic Contrastive Mixup
Carlos Manuel Ferreira Carvalho, Alberto Abad,
ICASSP, ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1-5, March 2025 - Memory-augmented conformer for improved end-to-end long-form ASR
Carlos Manuel Ferreira Carvalho, Alberto Abad,
INTERSPEECH, pages 2218-222, ISCA, August 2023 - TRIBUS: An end-to-end automatic speech recognition system for European Portuguese
Carlos Manuel Ferreira Carvalho, Alberto Abad,
IberSPEECH 2020, March 2021
Thesis (Author)
- End-to-end automatic speech recognition in Portuguese
Carlos Manuel Ferreira Carvalho, Instituto Superior Técnico, Universidade de Lisboa, Thesis, January 2021
Total Supervisions: 1
Master's Theses
- Exploring Memory Mechanisms in AI Agents for Graphical Interfaces
Stanislaw Talejko
Alberto Abad (advisor), Carlos Manuel Ferreira Carvalho (coadvisor)
Master's Thesis, Instituto Superior Técnico, Universidade de Lisboa, 2025-02-17 - Ongoing