About Me
Vivian Sedov, Ph.D. Candidate in Computer Vision
Royal Holloway University of London Email Me | GitHub | LinkedIn
"In the realm of Artificial Intelligence, there's no horizon but the one we create."
About Me
Graduating “summa cum laude” (1:1/First Class Honors) with a BSc in AI, I dove headfirst into the realms of deep learning, tirelessly optimizing and understanding different computer vision applications and unraveling the mysteries of reinforcement learning interlinked with contrastive networks. Serving as an engineer with Google Collab, I was immersed in cutting-edge technological challenges, pushing the boundaries of state-of-the-art algorithms.
Now, as a Ph.D. candidate, my dedication is channeled towards the nuanced intricacies of computer vision. My research probes the depths of reinforcement learning, marrying its adaptability with the precision of semantic segmentation to create models that not only ‘see’ but ‘understand’. Driven by the complexities of convolutional neural networks, adversarial attacks, and transfer learning, I’m on a mission to redefine the capabilities of machine vision.
Beyond just theory, my passion spills into real-world applications, from implementing robust object detection in cluttered scenes to architecting self-learning agents capable of navigating the digital and physical world. At the nexus of data science, mathematics, and algorithmic design, I stand committed to forging a brighter, AI-augmented future.
Research Contributions
During my B.Sc:
- GANs in Anime: An intricate analysis of generative abilities in Generative Adversarial Networks, spotlight on synthetic anime face generation.
- Vanishing Gradient & Model Collapse: Delved into the persistence of these issues and their impact on neural network convergence.
- ECG Analysis with Deep Learning: Comparative study on ECG data using RNNs and LSTM for heartbeat classification.
Interdisciplinary Interests
| Interest Area | Description |
|---|---|
| Aerospace Engineering & ML | Exploring federated learning & edge computing in satellite navigation systems. |
| Mathematical Optimization | Utilizing linear algebra & multivariable calculus for hyperparameter tuning and algorithm optimization. |
| Formula 1 & Automation | Championing reinforcement learning in simulating Formula 1 racing strategies. |
Current Endeavours
- Learning Contrastive Neural Networks.
- Crafting a Self-learning Chatbot.
- Mastering the art of Dataset and Model Creation.
- Venturing into Personal Research, Blogging, and unique Machine Learning Projects like HeartBeat Classification.
Past Achievements
- Authored my first research paper during my final year, which delved into advanced techniques in machine learning, receiving acclaim from leading researchers in the domain.
- Secured a position as an Assistant Researcher focusing on Interactive Visualisation of Disentangled Representations.
- Proud to have been an Ex-Google Collab Engineer during my tenure at the university, working closely with academia and industry professionals to bridge the gap between research and real-world applications.
- Successfully maintained and contributed to multiple open-source projects, advocating for transparency and community-driven development in AI.
- Spearheaded the development of a cutting-edge Drone Detection system using YoloV5.
- Designed and deployed a state-of-the-art Facial Recognition System.
- Embarked on my academic journey further by enrolling in a PhD program, with aspirations to unravel novel insights in the realm of computer vision and machine learning.
Currently Reading
| Book Name | Description |
|---|---|
| Deep Learning | By Ian Goodfellow, Yoshua Bengio & Aaron Courvile |
| Linear Algebra and Learning from Data | By Gilbert Strang |
| Deep Learning for Coders with Fastai | - |
| Genki 2 Japanese | - |
| Reinforcement Learning: An Introduction (Adaptive Computation and Machine Learning series) | By Richard S. Sutton, Andrew G. Barto, and Francis Bach. A comprehensive dive into reinforcement learning principles and applications. |
| Data Modeling with Snowflake: A practical guide to accelerating Snowflake development | By Serge Gershkovich. Foreword by Kent Graziano. An in-depth guide to Snowflake development using universal data modeling techniques. |
| Essential Math for AI: Next-Level Mathematics for Efficient and Successful AI Systems | By Hala Nelson. Explores advanced mathematical concepts critical to AI system optimization and success. |
| Image and Graphics: 12th International Conference, ICIG 2023, Nanjing, China, September 22–24, 2023 | Edited by Huchuan Lu, Wanli Ouyang, Hui Huang, Jiwen Lu, Risheng Liu, et al. Chronicles research findings presented at the ICIG 2023 conference. |
Let’s Connect!
Open for collaborations, discussions, or just a simple chat about the wonders of AI. 