Awarded a fellowship to work on a supervised research project in computational neuroscience
Supervisor: Dr Brendan Bicknell (UCL Häusser Lab & Gatsby Unit)
Project: Solving Sound Localisation using SNNs (part of the COMOB, initiative launched by Prof Dan Goodman) · GitHub repo currently private
Research outputs: poster for the in-person SURF Research Symposium hosted by the Simons Foundation (April 17, 2023, NYC) titled "Dendritic computations in spiking neural networks" + ongoing paper
Other activities: 1-month computational bootcamp · attended SNUFA 2022 conference · educational development fund to participate in the 3-week intensive Neuromatch Deep Learning course (July 2023)
Software Development Engineer I Intern in the Selection Contribution Platform organisation (3 months)
Learned about software engineering design · implemented a proof-of-concept (PoC) for an internal project sitting on the Amazon production traffic in Java (unit- & load-tested).
Artefacts: reviewed/merged code · technical documentations · final internal presentation to stakeholders
Skills: Java · Amazon Web Services · unit testing · dependency injection (Guice) · load shedding · technical writing (documentation) and presenting
AI Research Intern (Full-Time)
Imagination Technologies (Kings Langley)
07.2021 - 09.2021
Project: Individual research project within IMG AI Labs (3 months) on biologically plausible ANNs
Supervisors: James Imber, Sandra La Mantia, Tim Atherton
Experimented with biologically-inspired, Hebbian-like learning rules for feedforward multi-layer networks
Skills: PyTorch & Python for GPU-trained networks · Perforce version control · 20,000 words literature review (company material) · publicly available blog post reporting on some of my work at https://www.imaginationtech.com/blog/learning-from-the-best-how-nature-can-inspire-ai-research/
Undergraduate Research Assistant
Language & Cognition Lab (UCL Experimental Psych)
09.2020 - 05.2021
Coded the Ecological Language (ECOLANG) corpus collected by Prof Gabriella Viggliocco (PI): eye-gaze tracker/onomatopoeia/lexical affiliate/gesture/referent coding · manual tier alignment · word boundary correction · speech censoring/muting · reliability coding · transcription · Excel · Python scripting for automatic data preprocessing
Involved in the launch of the Human Simulation Paradigm study (study design, stimuli selection). Paper in process.
Skills: PRAAT · ELAN · Audacity · Excel · Python
Research Mentee in NLP
UCL Medical Physics & Biomedical Engineering
11.2020 - 04.2021
Project: "Developing a Segmentation Pipeline to Characterise and Quantify Radiomic Features of NSCLC CT Scans"
Worked on information extraction/retrieval in Electronic Health Records using Python, deep neural networks and NLP methods
Supervisor: Zihao Fang (PhD candidate)
Skills: Python (~3k LOC) · nltk · free-text unstructured data (data preprocessing & information extraction on EHRs) · Transformers · knowledge graphs · visualisations (word clouds, heatmaps) · literature review
Education
MSc Artificial Intelligence -
Imperial College London
09/2023 - 09/2024
MSci Psychology & Language Sciences -
University College London (UCL)
09/2019 - 06/2023
French Scientific (S) Baccalaureate -
Lycee Charles Peguy Paris (11e)
09/2016 - 06/2019
Skills
Python (incl OOP, numpy, pandas, scipy, matplotlib, seaborn, unit testing)
The Corporate Partnership Programme MSc Group Project Prize for excellence in the Group Project on the MSc AI Programme (£50) - 2024, Imperial College London.
James Martin Graduate Scholarship for the first year of the 1+3 DPhil (£9,000)- 2024, Keble College (Oxford).
Summer Extension to the SCGB Undergraduate Research Fellowship (SURF) in Computational Neuroscience (£6,000) - 2023, Simons Foundation & SCGB.
Educational Development Fund for the 3-week intensive Neuromatch Deep Learning Summer Course ($720) - 2023, Simons Foundation & SCGB.
SCGB Undergraduate Research Fellowship (SURF) in Computational Neuroscience ($25/h for 10h/w during 30 weeks) - 2022-2023, Simons Foundation & SCGB.
Nomination for the BPS Cognitive Psychology Undergraduate Project Prize - 2022, UCL.
Academic Excellence Scholarship (€800) for Summer School in Human AI - 2021, VU Amsterdam.
UCL Faculty Undergraduate Scholarship for Excellence (£3,000) x 2 - 2020 (nominated) & 2021 (awarded), UCL.
Dean's List x 2 in Years 2 & 3 - 2021 & 2022, UCL.
Jackson Lewis Scholarship for Best Year 2 Student in the Faculty of Brain Sciences (£200) - 2021, UCL.
Student Achievement Prizex 3for overall academic performance in Years 1 (£50), 3 (£100) and 4 (£100) - 2020/2022/2023, UCL.
Undergraduate Research Assistantship for the Language & Cognition Lab (£10/h for 1-20h/week) - 2020-2021, UCL.
Highest mark in the 2019 cohort and S section of the baccalaureate in my high school - 2019.
Publications
Published
Vellani, V., Zheng, S., Ercelik, D., & Sharot, T. (2023). The illusory truth effect leads to the spread of misinformation. Cognition, 236, 105421.
Ercelik, D. F., & Treur, J. (2021, September). A Controlled Adaptive Network Model for Joint Attention. In Biologically Inspired Cognitive Architectures Meeting (pp. 138-147). Cham: Springer International Publishing.
Blog Posts
Ercelik, D. F. (2021, October 4). Learning from the best: How nature can inspire AI research. Imagination Technologies Company Blog. https://www.imaginationtech.com/blog/learning-from-the-best-how-nature-can-inspire-ai-research/
Ercelik, D. F. (2023, April 4). What deaf people use their “auditory” brain regions for: A study of non-auditory cognition in deafness. UCL PALS Blog. https://reflect.ucl.ac.uk/pals-msci-placement-projects/2023/04/05/what_deaf_people_use_their_auditory_brain_regions_for-a_study_of_non-auditory_cognition_in_deafness/
In Progress
Spiking neural network models of sound localisation via a massively collaborative process [preprint] (Gosh et al., July 2024): DOI: 10.1101/2024.07.19.604252
Examining the effectiveness of multimodal cues using the Human Simulation Paradigm (HSP) (Khachatoorian et al., TBA)
Languages
French (Native) · Turkish (Native) · English (C1 - IELTS Band 8) · German (B1)