Quantum Machine Learning: I recently encountered the fascinating intersection of quantum computing and computer vision, specifically the concept of quantum neural networks. This intriguing topic led me to dive deeper into quantum computing, exploring its principles, applications, and potential in modern AI systems. To better understand this field, I studied foundational concepts in quantum mechanics and quantum computing, which culminated in the creation of a slide presentation summarizing my findings. This exploration not only broadened my perspective on how quantum computing could revolutionize machine learning but also sparked a deeper interest in its applications across various domains, including computer vision. 

Slide: Quantum Computing for the very curious 

Single Cell Foundational Models: I’ve always thought large language models (LLMs) were primarily useful for text and image data. However, the idea that transformers, a technique I associate with text generation, could be applied to biological fields, particularly in single-cell multi-omics, was initially intimidating to me. But after reading the paper on scGPT: Towards Building a Foundation Model for Single-Cell Multi-omics Using Generative AI, which explores how generative pre-trained transformers can be leveraged to analyze complex biological data, I realized the incredible potential of LLMs in diverse domains beyond traditional applications. This insight opened my eyes to the possibility of using machine learning and transformers to revolutionize industries like healthcare, drug discovery, and material science. Based on my reading, I created a slide summarizing the key concepts of scGPT, focusing on its ability to optimize tasks such as cell-type annotation, perturbation prediction, and multi-omic integration. The paper highlights scGPT’s approach to understanding cellular data and how it serves as a foundational model for single-cell analysis, providing state-of-the-art performance and reducing the need for task-specific methods.

Slide: SC-GPT