Gonçalo Mordido

Senior AI Researcher

Huawei Ireland Research Centre

goncalomordido [at] gmail [dot] com

Publications

2024

Exploring quantization for efficient pre-training of Transformer language models.

K. Chitsaz, Q. Fournier, G. Mordido, S. Chandar.

Findings of EMNLP 2024

Should we attend more or less? Modulating attention for fairness.

A. Zayed, G. Mordido, S. Shabanian, S. Chandar.

COLM 2024

Promoting exploration in memory-augmented Adam using critical momenta.

P. Malviya, G. Mordido, A. Baratin, R. Harikandeh, J. Huang, S. Lacoste-Julien, R. Pascanu, S. Chandar.

Transactions on Machine Learning Research

Why don’t prompt-based fairness metrics correlate?

A. Zayed, G. Mordido, I. Baldini, S. Chandar.

ACL 2024

Lookbehind-SAM: k steps back, 1 step forward.

G. Mordido, P. Malviya, A. Baratin, S. Chandar.

ICML 2024

Fast and accurate output error estimation for memristor-based deep neural networks.

J. Kern, S. Henwood, G. Mordido, E. Dupraz, A. Bey, Y. Savaria, F. Leduc-Primeau.

IEEE Transactions on Signal Processing

Fairness-aware structured pruning in Transformers.

A. Zayed, G. Mordido, S. Shabanian, I. Baldini, S. Chandar.

AAAI 2024

2023

Training DNNs resilient to adversarial and random bit-flips by learning quantization ranges.

K. Chitsaz, G. Mordido, J. David, F. Leduc-Primeau.

Transactions on Machine Learning Research

Deep learning on a healthy data diet: Finding important examples for fairness and performance.

A. Zayed, P. Parthasarathi, G. Mordido, H. Palangi, S. Shabanian, S. Chandar.

AAAI 2023

2022

Sharpness-aware training for accurate inference on noisy DNN accelerators.

G. Mordido, S. Chandar, F. Leduc-Primeau.

CoLLAs 2022 workshop & Edge Intelligence workshop 2022

Improving meta-learning generalization with activation-based early-stopping.

S. Guiroy, C. Pal, G. Mordido, S. Chandar.

CoLLAs 2022

MemSE: Fast MSE prediction for noisy memristor-based DNN accelerators.

J. Kern, S. Henwood, G. Mordido, E. Dupraz, A. Aissa-El-Bye, Y. Savaria, F. Leduc-Primeau.

AICAS 2022

Tiny CNN for seizure prediction in wearable biomedical devices.

Y. Zhang, Y. Savaria, S. Zhao, G. Mordido, M. Sawan, F. Leduc-Primeau.

EMBC 2022

2021

Compressing 1D time-channel separable convolutions using sparse random ternary matrices.

G. Mordido, M. Keirsbilck, A. Keller.

INTERSPEECH 2021

Assessing image and text generation with topological analysis and fuzzy logic.

G. Mordido*, J. Niedermeier*, C. Meinel.

WACV 2021

2020

Mark-Evaluate: Assessing language generation using population estimation methods.

G. Mordido, C. Meinel.

COLING 2020

Best student forcing: A simple training mechanism in adversarial language generation.

J. Sauder*, T. Hu*, X. Che, G. Mordido, H. Yang and C. Meinel.

LREC 2020

Monte Carlo gradient quantization.

G. Mordido, M. Keirsbilck, A. Keller.

CVPR 2020 EDLCV workshop

Improving the evaluation of generative models with fuzzy logic.

J. Niedermeier*, G. Mordido* and C. Meinel.

AAAI 2020 Meta-Eval workshop

microbatchGAN: Stimulating diversity with multi-adversarial discrimination.

G. Mordido, H. Yang, and C. Meinel.

WACV 2020

2019

Instant quantization of neural networks using Monte Carlo methods.

G. Mordido*, M. Keirsbilck*, A. Keller.

NeurIPS 2019 EMC2 workshop

2018

Pseudo-ground-truth for adversarial text generation using reinforcement learning.

J. Sauder, X. Che, G. Mordido, H. Yang and C. Meinel.

NeurIPS 2018 Deep RL workshop

Dropout-GAN: Learning from a dynamic ensemble of discriminators.

G. Mordido, H. Yang, and C. Meinel.

KDD 2018 Deep Learning Day

2017

Automatic organisation, segmentation, and filtering of user-generated audio content.

G. Mordido, J. Magalhaes, and S. Cavaco.

MMSP 2017

Automatic organisation and quality analysis of user-generated content with audio fingerprinting.

G. Mordido, J. Magalhaes, and S. Cavaco.

EUSIPCO 2017

Patents

2022

Incorporating a ternary matrix into a neural network.

A. Keller, G. Mordido, M. Van keirsbilck.

US Patent

2019

Representing a neural net utilizing paths within the network to improve a performance of the neural net.

A. Keller, G. Mordido, N. Gamboa, M. Van keirsbilck.

US Patent