Haytham

Artificial Intelligence / Machine Learning.
Royal Melbourne Institute of Technology (RMIT).
Alumnus: Facebook Research, Oculus Research, & WorleyParsons.
Google Scholar | Github | LinkedIn | Twitter


Current



Call for Papers for Neural Networks Special Issue on Lifelong Learning due 28 Feb 2022.
I am looking for strong students interested in machine learning, so please get in touch if you would like to work with me at RMIT — competitive scholarships available.


Recent



Oct 2021: Paper: Knowledge capture and replay for continual learning, accepted at WACV’22.
Sep 2021: Paper: Recent developments in the implementation of a bidirectional LSTM deep neural network for aircraft operational loads monitoring, accepted at SciTech’22.
Sep 2021: Paper: Data-driven flight load prediction using modal decomposition techniques, accepted at SciTech’22.
Aug 2021: Call for Papers for Neural Networks Special Issue on Lifelong Learning due 28 Feb 2022.
Jul 2021: Patent: Systems and methods for hearing assessment and audio adjustment published.
Jul 2021: Teaching COSC2959/COSC2960 Foundations of Artificial Intelligence, Jul’21.
Jul 2021: Our Special Issue on Lifelong Learning was accepted in Neural Networks.
May 2021: Paper: Deep learning airframe load prediction: A data-driven system for aircraft structural health management, accepted at AIAC’21.
May 2021: Paper: Buffet load prediction via frequency response functions, accepted at AIAC’21.
Mar 2021: Teaching COSC2676/COSC2752 Programming Fundamentals for Scientists, Mar’21.
Jan 2021: At IJCAI’20, Virtual, to present Large scale audiovisual learning of sounds with weakly labeled data.
Previous: Previous news.


About



I am on the faculty of the School of Computing Technologies at the Royal Melbourne Institute of Technology (RMIT), Melbourne, VIC, where I am affiliated with the Evolutionary Computing and Machine Learning (ECML) Group, the AI Innovation Lab, and the Centre for Information Discovery and Data Analytics (CIDDA).

I was a Postdoctoral Research Scientist at Facebook Research in Seattle, WA, from August 2018 to January 2020. I received a PhD from RMIT in 2019. My PhD thesis is titled, Continual Deep Learning via Progressive Learning. During my PhD, I was a Research Intern at Facebook with Facebook Reality Labs (Oculus VR Research) and Facebook AI Research (FAIR). I was also a Teaching Assistant and Guest Lecturer at RMIT. Formerly, I received an MSc (Research) and a BEng (Hons) in Electrical and Electronics Engineering from Petronas University in 2014 and 2012 respectively, and worked as an Electrical/Electronics Engineer in the engineering consulting industry for three years.


Research



My research interests are broadly in artificial intelligence, machine learning, deep learning, and machine perception. I am primarily interested in learning systems that systematically generalize from limited labelled data.


Publications



Papers

Saisubramaniam Gopalakrishnan, Pranshu Ranjan Singh, Haytham M. Fayek, Savitha Ramasamy, and Arulmurugan Ambikapathi.
Knowledge capture and replay for continual learning.
In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, USA, Jan 2022.
arXiv pdf bib

Michael Candon, Haytham M. Fayek, Oleg Levinski, Stephan Koschel, and Pier Marzocca.
Recent developments in the implementation of a bidirectional LSTM deep neural network for aircraft operational loads monitoring.
In AIAA SciTech Forum, San Diego, USA, Jan 2022.

Stephan Koschel, Robert Carrese, Michael Candon, Haytham M. Fayek, Pier Marzocca, and Oleg Levinski.
Data-driven flight load prediction using modal decomposition techniques.
In AIAA SciTech Forum, San Diego, USA, Jan 2022.

Haytham M. Fayek, Michael Candon, Oleg Levinski, Stephan Koschel, and Pier Marzocca
Deep learning airframe load prediction: A data-driven system for aircraft structural health management.
In 19th Australian International Aerospace Congress (AIAC), Melbourne, Australia, Nov 2021.

Stephan Koschel, Robert Carrese, Haytham M. Fayek, Pier Marzocca, and Oleg Levinski.
Buffet load prediction via frequency response functions.
In 19th Australian International Aerospace Congress (AIAC), Melbourne, Australia, Nov 2021.

Haytham M. Fayek and Justin Johnson.
Temporal reasoning via audio question answering.
IEEE/ACM Transactions on Audio, Speech, and Language Processing, Jul 2020.
link arXiv code bib

Haytham M. Fayek and Anurag Kumar.
Large scale audiovisual learning of sounds with weakly labeled data.
In 29th International Joint Conference on Artificial Intelligence (IJCAI), Yokohama, Japan, Jul 2020.
link pdf arXiv bib

Haytham M. Fayek, Lawrence Cavedon, and Hong Ren Wu.
Progressive learning: A deep learning framework for continual learning.
Neural Networks, vol. 128, pp. 345–357, May 2020.
link bib

Sharath Adavanne, Haytham M. Fayek, and Vladimir Tourbabin.
Sound event classification and detection with weakly labeled data.
In 4th Workshop on Detection and Classification of Acoustic Scenes and Events (DCASE), New York, USA, pp. 15–19, Oct 2019.
pdf bib

Haytham M. Fayek, Lawrence Cavedon, and Hong Ren Wu.
On the transferability of representations in neural networks between datasets and tasks.
In Continual Learning Workshop, 32nd Neural Information Processing Systems (NeurIPS), Montréal, Canada, Dec 2018.
pdf arXiv bib

Haytham M. Fayek.
MatDL: A lightweight deep learning library in MATLAB.
The Journal of Open Source Software, vol. 2, no. 19, pp. 413, Nov 2017.
link pdf code bib

Haytham M. Fayek, Laurens van der Maaten, Griffin D. Romigh, and Ravish Mehra.
On data-driven approaches to head-related transfer function personalization.
In Audio Engineering Society (AES) Convention 143, New York, USA, Oct 2017.
link pdf bib

Haytham M. Fayek, Margaret Lech, and Lawrence Cavedon.
Evaluating deep learning architectures for speech emotion recognition.
Neural Networks, vol. 92, pp. 60–68, Aug 2017.
link bib

Haytham M. Fayek.
A deep learning framework for hybrid linguistic-paralinguistic speech systems.
In 2nd Doctoral Consortium at Interspeech, Berkeley, USA, Sep 2016.
link pdf bib

Haytham M. Fayek, Margaret Lech, and Lawrence Cavedon.
On the correlation and transferability of features between automatic speech recognition and speech emotion recognition.
In 17th Annual Conference of the International Speech Communication Association (Interspeech), San Francisco, USA, pp. 3618–3622, Sep 2016.
link pdf slides bib

Haytham M. Fayek, Margaret Lech, and Lawrence Cavedon.
Modeling subjectiveness in emotion recognition with deep neural networks: Ensembles vs soft labels.
In International Joint Conference on Neural Networks (IJCNN), Vancouver, Canada, pp. 566–570, Jul 2016.
link poster bib

Haytham M. Fayek, Margaret Lech, and Lawrence Cavedon.
Towards real-time speech emotion recognition using deep neural networks.
In 9th International Conference on Signal Processing and Communication Systems (ICSPCS), Cairns, Australia, pp. 1–5, Dec 2015.
link bib

P. N. Q. Nhon, Irraivan Elamvazuthi, Haytham M. Fayek, S. Parasuraman, and M.K.A. Ahamed Khan.
Intelligent control of rehabilitation robot: Auto tuning PID controller with interval type 2 fuzzy for DC servomotor.
Procedia Computer Science, vol. 42, Medical and Rehabilitation Robotics and Instrumentation (MRRI2013), pp. 183–190, Dec 2014.
link bib

Haytham M. Fayek, Irraivan Elamvazuthi, N. Perumal, and Bala Venkatesh.
A controller based on optimal type-2 fuzzy logic: Systematic design, optimization and real-time implementation.
ISA Transactions, vol. 53(5), pp. 1583–1591, Sep 2014.
link bib

Haytham M. Fayek, Irraivan Elamvazuthi, N. Perumal, and Bala Venkatesh.
The impact of DFIG and FSIG wind farms on the small signal stability of a power system.
In 5th International Conference on Intelligent and Advanced Systems (ICIAS), Kuala Lumpur, Malaysia, pp. 1–6, Jun 2014.
link slides bib

Haytham M. Fayek and Irraivan Elamvazuthi.
Real-time implementation of a type-2 fuzzy logic controller to control a DC servomotor with different defuzzification methods.
In 18th International Conference on Methods and Models in Automation and Robotics (MMAR), Miedzyzdroje, Poland, pp. 86–91, Aug 2013.
link bib (Runner up Best Young Author)

Haytham M. Fayek and Irraivan Elamvazuthi.
Type-2 fuzzy logic PI (T2FLPI) based DC servomotor control.
Journal of Applied Sciences Research, vol. 8(5), pp. 2564–-2574, 2012.
bib

Theses

Haytham M. Fayek.
Continual deep learning via progressive learning.
PhD Thesis, RMIT University, 2019.
link pdf bib

Haytham M. Fayek.
Systematic design of optimal type II fuzzy logic controllers with applications to wind power.
MSc Thesis, Petronas University, 2014.
bib

Haytham M. Fayek.
Fuzzy logic based motor control.
BEng Thesis, Petronas University, 2012.
bib

Patent

Haytham M. Fayek A. A. Abokela and Antonio John Miller.
Systems and methods for hearing assessment and audio adjustment.
US Patent App. 16/745,287, Jul 2021.
link pdf bib

Irraivan Elamvazuthi and Haytham M. Fayek.
Method for controlling a high speed DC servomotor that controls a robotic arm for the PCB industry (FT2RC).
Patent Filing No. PI2012700978, Nov 2012.
bib


Talks



Invited Talks

Progressive Learning.
A*Star.
Singapore, Aug 2020.

Linguistic and Paralinguistic Speech Recognition.
Voicea.
Menlo Park, USA, Dec 2017.

Learning from Data and Prior Knowledge.
Facebook.
Redmond, USA, Oct 2017.

Will Deep Learning Lead to AI?
Swinburne University.
Melbourne, Australia, Sep 2017.

Will Deep Learning Lead to AI?
Melbourne Machine Learning & AI Meetup.
Melbourne, Australia, Aug 2017.
slides link

The Types of Power Quality Problems and their Effects on Electrical & Electronic Systems.
Fire Protection Association of Malaysia (FPAM).
Kuala Lumpur, Malaysia, Mar 2014.

Conference Presentations

Large Scale Audiovisual Learning of Sounds with Weakly Labeled Data.
29th International Joint Conference on Artificial Intelligence (IJCAI).
Yokohama, Japan, Jan 2021.

On the Transferability of Representations in Neural Networks Between Datasets and Tasks.
Continual Learning Workshop, 32nd Neural Information Processing Systems (NeurIPS 2018).
Montréal, Canada, Dec 2018.
poster

On Data-driven Approaches to Head-related Transfer Function Personalization.
Audio Engineering Society (AES) Convention 143.
New York, USA, Oct 2017.

On the Correlation and Transferability of Features between Automatic Speech Recognition and Speech Emotion Recognition.
17th Annual Conference of the International Speech Communication Association (Interspeech).
San Francisco, USA, Sep 2016.
slides

A Deep Learning Framework for Hybrid Linguistic-Paralinguistic Speech Systems.
2nd Interspeech Doctoral Consortium.
Berkeley, USA, Sep 2016.

Modeling Subjectiveness in Emotion Recognition with Deep Neural Networks: Ensembles vs Soft Labels.
29th International Joint Conference on Neural Networks (IJCNN).
Vancouver, Canada, Jul 2016.
poster

Towards Real-time Speech Emotion Recognition using Deep Neural Networks.
9th International Conference on Signal Processing and Communication Systems (ICSPCS).
Cairns, Australia, Dec 2015.

The Impact of DFIG and FSIG Wind Farms on the Small Signal Stability of a Power System.
5th International Conference on Intelligent and Advanced Systems (ICIAS).
Kuala Lumpur, Malaysia, Jun 2014.
slides

Real-time Implementation of a Type-2 Fuzzy Logic Controller to Control a DC Servomotor with Different Defuzzification Methods.
18th International Conference on Methods & Models in Automation & Robotics (MMAR).
Miedzyzdroje, Poland, Aug 2013.

Panel Discussions

Talking Computers: How Voice Technology is Changing the Human Machine Interaction.
ACMI X, Academy Xi, and FAB9.
Melbourne, Australia, Jun 2018.
link


Research Students



Current

Emma Pretty, Ph.D. Student, RMIT, 2021–Present.
Fitrio Pakana, M.Sc. Student, RMIT, 2021–Present.
Halide Göknur Aydoğan, Ph.D. Student, RMIT, 2021–Present.
Nermine Hendy, Ph.D. Student, RMIT, 2021–Present.
Yameng Peng, Ph.D. Student, RMIT, 2020–Present.

Alumni

Lior Madmoni, Research Intern, Facebook Reality Labs, 2019. (Ben-Gurion University)
Shengjie Bi, Research Intern, Facebook Reality Labs, 2019. (Dartmouth College)
Sharath Adavanne, Research Intern, Facebook Reality Labs, 2018. (Tampere University)


Teaching



Courses

COSC2959/COSC2960 Foundations of Artificial Intelligence.
RMIT, Australia, July 21.

COSC2676/COSC2752 Programming Fundamentals for Scientists.
RMIT, Australia, March 21.

COSC2676/COSC2752 Programming Fundamentals for Scientists.
RMIT, Australia, July 20.

COSC2676/COSC2752 Programming Fundamentals for Scientists.
RMIT, Australia, March 20.

Guest Lectures

Digital Image Watermarking.
EEET2169/EEE1255 Image Processing / Image Systems Engineering.
RMIT, Australia, April 2018.

Sampling and Reconstruction.
EEET2113 Signals and Systems II.
RMIT, Australia, March 2016.

Signals in Noise.
EEET2113 Digital Signal Processing.
RMIT, Australia, April 2015.


Contact



Dr Haytham Fayek
School of Computing Technologies
Royal Melbourne Institute of Technology (RMIT)
Building 14, Level 11, Room 03
124 La Trobe Street, Melbourne VIC 3000
Australia

Email:
Phone:
Postal: