Mert Kilickaya


I am a computer vision and machine learning researcher. My goal is to build autonomous visual learners that can self-improve with minimal human intervention: How can we learn from uncurated and unlabeled data?

I work at the Learning to Learn Lab, where I focus on self-supervision and continual learning. Before that, I obtained my PhD at the QUvA Lab , where I tackled visual detection using limited supervision, under the guidance of Arnold Smeulders.

When I'm not researching, I love capturing beautiful cities with my GoPro. Tavira, Portugal, and Montreal, Canada, are among my favorite places. If you have any suggestions for my next adventure, feel free to contact me!

Contact Me

Email  |  CV  |  Google Scholar  |  Twitter  |  LinkedIn



TU Eindhoven
Learning to Learn Lab, Eindhoven University of Technology, Netherlands

Post-doc Researcher

QUvA Deep Vision Lab
QUvA Deep Vision Lab, University of Amsterdam, Netherlands

PhD Researcher

Huawei Visual Search Lab
Huawei Visual Search Lab, Helsinki, Finland
2021 March - 2022 January

Research Scientist Intern

Université Laval
Université Laval Computer Vision Lab, Quebec, Canada
2015 March - 2015 October

Research Intern


My research goal is to reduce the human supervision need of (continual) visual learners.

PontTuset Towards Label-Efficient Incremental Learning: A Survey
Mert Kilickaya, Joost van de Weijer, Yuki Asano
Preprint 2023
arXiv / Slides / Github

We survey semi-, few-shot and self-supervised incremental learning.

PontTuset Are Labels Needed for Incremental Instance Learning?
Mert Kilickaya, Joaquin Vanschoren
CVPRW 2023 (Oral)
arXiv / Slides

We introduce VINIL, a self-supervised incremental instance learner.

PontTuset Human-Object Interaction Detection via Weak Supervision
Mert Kilickaya, Arnold Smeulders
BMVC 2021
arXiv / Slides

We introduce Align-Former to detect human-object interactions from an image.

PontTuset Structured Visual Search via Composition-aware Learning
Mert Kilickaya, Arnold Smeulders
WACV 2021
arXiv / Slides

We introduce composition-aware learning to improve visual image search.

PontTuset Re-evaluating Automatic Metrics for Image Captioning
Mert Kilickaya, Aykut Erdem, Nazli Ikizler-Cinbis, Erkut Erdem
EACL 2017
arXiv / Slides

We propose Word-Mover Distance to evaluation image captioning.


Visual Image Search via Conversational Interaction (Huawei)

Mert Kilickaya, Baiqiang XIA

US Patent, 2022, Link

Network for Interacted Object Localization (Qualcomm)

Mert Kilickaya, Arnold Smeulders

US Patent, 2022, Link

Subject-Object Interaction Recognition Model (Qualcomm)

Mert Kilickaya, Stratis Gavves, Arnold Smeulders

US Patent, 2022, Link

Context-driven Learning of Human-object Interactions (Qualcomm)

Mert Kilickaya, Noureldien Hussein, Stratis Gavves, Arnold Smeulders

US Patent, 2021, Link


I am very fortunate to cross roads with the kind souls below.

Student  Picture Fangqin Zhou (PhD, TU/e) 2022- Self-Supervised Learning for AI4Good
Student  Picture Ran Piao (MSc, TU/e) 2022- Rapid ViT Architecture Search without Training
Student Picture Ceren Gok (PhD, TU/e) 2022-2023 Learning to Adapt for Continual Learning
Student Picture Kishan Parshotam (MSc, UvA) 2020-2021 Continual Learning of Object Instances
Student Picture Tarun Krishna (MSc, UvA) 2019-2020 Disentangling Rotation from Identity for Visual Instance Search

Source code credit to Dr. Jon Barron and ChatGPT