Postdoctoral Research Fellow - Computer vision, AI, machine learning (2 years)

King Abdullah University of Science and Technology: Postdoc Positions:

Computer, Electrical and Mathematical Science and Engineering Division (postdoc)

Location

KAUST, Thuwal 23955, KSA

Open Date

Oct 02, 2021

Description

The Computer Vision-Core Artificial Intelligence Research (Vision-CAIR) group led by Prof. Mohamed Elhoseiny at the CS Program of the King Abdullah University of Science and Technology (KAUST) is looking for postdoctoral researchers in the area of computer vision, AI, and machine learning. The initial appointment will be for 2 years with a possible extension with a tentative start date in January 2022 but may also start earlier. KAUST offers competitive/generous postdoc compensations.

The Research Vision-CAIR group performs research and develops computational approaches in the following research themes:

( a) learning efficiency, computational creativity (zero, few-shot, and long-tail learning of 2D and 3D vision tasks. This also includes efficient generative models that are capable of generating and understanding unseen art and fashion in 2D and 3D);

(b) continual learning (e.g., alleviating catastrophic forgetting in various learning settings including recognition RL);

(c) vision and language (this overlaps with the former themes as 2D and 3D vision is often integrated with language).

Related papers by the Vision-CAIR members

Theme A :

Aligning Latent and Image Spaces to Connect the Unconnectable ICCV 2021, Ivan Skorokhodov, Grigory Sotnikov, Mohamed Elhoseiny

Adversarial Generation of Continuous Images, Ivan Skorokhodo, Savva Ignatyev, Mohamed Elhoseiny, CVPR, 2021 ,

Class Normalization for (Continual)? Zero-Shot Learning, Ivan Skorokhodov, Mohamed Elhoseiny, ICLR , 2021

Exploring Long Tail Visual Relationship Recognition with Large Vocabulary, ICCV , 2021, Sherif Abdelkarim, Aniket Agarwal, Panos Achlioptas, Jun Chen, Jiaji Huang, Boyang Li, Kenneth Church, Mohamed Elhoseiny

Wölfflin's Affective Generative Analysis of Artworks, Divyansh Jha, Hanna Chang, and Mohamed Elhoseiny, ICCC , 2021,

Ahmed Elgammal, Bingchen Liu, Mohamed Elhoseiny, and Marian Mazzone, C AN: Creative Adversarial Networks, International Conference on Computational Creativity( ICCC ), 2017

Theme B continual learning (selected papers):

Class Normalization for (Continual)? Zero-Shot Learning, Ivan Skorokhodov, Mohamed Elhoseiny, ICLR , 2021

Compositional Continual Language Learning , ( ICLR '20), Yuanpeng Li, Liang Zhao, Ken Church, Mohamed Elhoseiny, Code: https: // github.com/yli1/CLCL

Uncertainty-guided Continual Learning with Bayesian Neural Networks ( ICLR'20 ), Sayna Ebrahimi, Mohamed Elhoseiny, Trevor Darrell, Marcus Rohrbach, Code: https: // github.com/SaynaEbrahimi/UCB

Efficient Lifelong Learning with A-GEM ( ICLR '19) Arslan Chaudhry, Marc'Aurelio Ranzato, Marcus Rohrbach, Mohamed Elhoseiny https: // github.com/facebookresearch/agem

Memory Aware Synapses: Learning what (not) to forget. ( ECCV '18) R Aljundi, F Babiloni, M Elhoseiny, M Rohrbach and T Tuytelaars Code: https: // github.com/rahafaljundi/MAS-Memory-Aware-Synapses

Theme C Vision &Language learning (selected papers):

VisualGPT: Data-efficient Image Captioning by Balancing Visual Input and Linguistic Knowledge from Pretraining, Jun Chen, Han Guo, Kai Yi, Boyang Li, Mohamed Elhoseiny, Arxiv, 2021, Vision& Language, Transformers

CIZSL++: Creativity Inspired Generative Zero-Shot Learning Mohamed Elhoseiny, Kai Yi, Mohamed Elfeki, TPAMI 2021, submission, Zero-shot Learning, Generative Models.

ArtEmis: Affective Language for Art Panos Achlioptas, Maks Ovsjanikov, Kilichbek Haydarov, Mohamed Elhoseiny, Leonidas Guibas, CVPR , 2021.

Mohamed Elhoseiny, Mohamed Elfeki Creativity Inspired Zero Shot Learning, Thirty-sixth International Conference on Computer Vision ( ICCV ), 2019

Videos

A video explaining what is Imagination-Inspired AI relating to themes (a) and (c) is available here:

https: // www. youtube.com/watch?v=Z1dytXfbFkM https: // webcast.kaust.edu.sa/Mediasite/Showcase/default/Presentation/8e74a8cf69384ca3bbf4b88bc2addea51d

artemisdataset.org (Affective Language for Visual Art) at the intersection between Art and Language & Vision

A presentation covering continual learning methods relating to theme (b) is available here:

https: // www. dropbox.com/s/sdtbmptu7a2chks/deepcontinuallearning.pdf

More information about the research group can be found here:

https: // cemse.kaust.edu.sa/vision-cair

https: // twitter.com/KAUSTVisionCAIR

https:// www. mohamed-elhoseiny.com/

Requirements include good communication (verbal, writing, etc) and excellent publication record in high-quality journals and/or conference proceedings such as CVPR, ICCV, NeurIPS, ICML, ICLR, ECCV, Nature, TPAMI, AAAI, Science, etc. It is an opportunity to develop core algorithmic advances and aim at publishing them at these top venues. More details about the position can be found here.

As a KAUST postdoc or researcher, you will have access to state-of-the-art research facilities, including a supercomputer. You will work in an exciting international environment with researchers from over 100 nationalities and will have the opportunity to participate in the development of computational algorithms and tools for problems of relevance to Vision-CAIR research group and Visual Computing Center (VCC) with a particular focus on learning efficiency, computational creativity, continual learning, computer vision, and natural language processing.

You will have the opportunity to boost your career development and training, read more here. Post-docs receive highly competitive salaries and other benefits, inclusive of accommodation, healthcare, medical and dental insurance, relocation allowance, one round-trip travel to home country per year, access to on-campus daycare and International Baccalaureate (IB) school, campus transportation (on-demand as well as scheduled), exceptional community services, facilities, programs, and events to help you enjoy an active, healthy, balanced and fulfilled life at KAUST. These community resources allow postdocs and researchers to balance their professional and family commitments.

Qualifications

We welcome candidates with a PhD in Computer Science, Engineering, Informatics, or related areas, with a strong background in the theory and practice of one or more of following research areas - machine learning efficiency, computational creativity, continual learning, computer vision, and natural language processing.

We require documented experience about computer vision, machine learning, deep learning or optimization, demonstrated by high-quality publications. We consider as an advantage documented experience publishing in top Computer Vision and AI conferences including CVPR/ICCV/ECCV/ICLR/ICML/NeurIPS/AAAI. We consider as an added advantage documented experience in specific instances of following topics: creative AI, zero-shot learning, generative models, and continual learning.

The call is open to all applicants from any nationality. International applications are strongly encouraged. Fluent written and verbal communication skills in English are mandatory. We require commitment, team working and a critical mind.

Application Instructions

You will be required to complete an application form and upload all of the following materials:

Curriculum Vitae,

publications especially in CVPR/ICCV/ECCV/ICLR/ICML/NeurIPS/AAAI/TPAMI.

Copies of up to three papers substantiating your background and experience in these topics.

Contact information for 3 referees.

For more information, please contact:

Prof. Mohamed H. Elhoseiny

via email at: [email protected],

Subject: COMPUTER VISION - POSTDOCTORAL APPLICATIONS.

Next cut-off date for applications: November 30, 2021.

Please click on the "Apply Now" button to submit your application.