Welcome to my GitHub profile! I am currently pursuing my Bachelor’s degree in Computer Science at the University of Tehran, with a deep focus on Artificial Intelligence and Machine Learning.
- 🔭 Current Focus: Transformer-based time series analysis for cardiovascular disease diagnosis and deepening my knowledge in leading-edge AI methods.
- 🎓 Academic Goals: Seeking PhD opportunities to contribute to cutting-edge AI research, particularly in generative AI, Human-AI Collaboration, and Agentic AI.
- 🌱 Always Learning: Continuously exploring new technologies and frameworks to drive innovation.
- 🛠️ Tools I use:
- Artificial Intelligence (AI)
- Generative AI and Foundation Models
- Multimodal AI (language, time series, vision)
- Human-AI Collaboration and Agentic AI
- Applications of AI to Healthcare
- Neuro-inspired Machine/Deep Learning
Supervised by Dr. Babaali | In Progress
University of Tehran, Faculty of Science, Department of Mathematics, Statistics, and Computer Science
- Objective: Leverage and extend transformer models for ECG Time Series Classification (TSC), with a focus on improved accuracy for arrhythmia detection.
- Approach: Investigate transformer architectures to capture both local and long-range dependencies in ECG data for comprehensive feature extraction. Evaluate generated features for arrhythmia detection and other potential machine learning based diagnostic tools for heart disease.
Hosseini, D., & Amini, A. (2024) | Supervised by Dr. G Rokni Lamouki
University of Tehran, Faculty of Science, Department of Mathematics, Statistics, and Computer Science
- Examined neuron dynamics and their simulation in discrete time, focusing on three models: Leaky Integrate-and-Fire (LIF), Exponential Leaky Integrate-and-Fire (ELIF), and Adaptive Exponential Leaky Integrate-and-Fire (AELIF).
- Findings highlighted the applicability of each model and the largest acceptable time step in simulation.
GitHub Repository | Spring 2024
- Technologies: Python, PyTorch, CoNeX
- Developed a spiking neural network using CoNeX, integrating multi-scale Gabor and DoG filters. Explored different filter sizes to analyze what features are learned by the CNN model. Applied the model to both synthetic and real-world images, replicating biological visual processing.
GitHub Repository | Fall 2023
- Technologies: Python, PyTorch, Scikit-learn, Numpy
- Developed a deep learning U-Net model using the Kvasir-SEG dataset for automatic polyp detection. Achieved high accuracy and performance metrics, enabling early cancer screening. Optimized hyperparameters and used data augmentation for better clinical practices.
GitHub Repository | Fall 2023
- Technologies: Python, PyTorch, Scikit-learn, Pretrained transformers
- Scraped Persian literature websites to create a dataset containing 10 authors with 30 documents each. Developed a BERT-based machine learning model for authorship identification.
GitHub Repository | Spring 2023
- Technologies: Python, Scikit-Learn, Numpy
- Developed and implemented bio-inspired AI algorithms, like genetic algorithms and simulated annealing, to theoretically model and solve cutting stock problems, leading to improved resource utilization and industry-specific solutions.
GitHub Repository | Spring 2023
- Technologies: Assembly (x86 Nasm), Text Processing
- Developed a text editor in Assembly to read, display, and edit file content, with features for file statistics (characters, words, lines) and search-and-replace functionality.
GitHub Repository | Spring 2022
- Technologies: Python, PyTorch, CoNeX
- Implemented various pretrained and custom CNN models for grapevine leaves classification. Used data augmentation with a data generator for better results and designed an autoencoder model for dimensionality reduction.
Due to confidentiality agreements with the companies for which these projects were developed, the code is not accessible.
- Music Recommendation System
- Chatbot
- Admin Panel
- OCR for Text Recognition
- Recognizing Tooth Issues Based on Images
- Persian - native
- English - fluent (C1, IELTS: 7.5)
Looking forward to connecting with like-minded individuals and contributing to exciting projects!
I'm always open to collaboration, discussion, and new opportunities. Feel free to reach out to me through:
- Primary email: [email protected] 📧
- Secondary email: [email protected] 📨
- LinkedIn: AmirAli Amini