PhD Lite

The Significance of Null Results in Physics Education Research

This podcast discusses the significance of null results in scientific research, particularly in physics education. It argues that null findings, often overlooked, provide crucial information about the limitations of theories and methods.
Paper: https://arxiv.org/pdf/1810.10071

Decompose the Model: Mechanistic Interpretability in Image Models with Generalized Integrated Gradients (GIG)

This conversation summarizes a research paper introducing Generalized Integrated Gradients (GIG) for interpreting image models. GIG analyzes the entire dataset, unlike previous methods focusing on individual classes, to identify shared concepts across images.Paper: https://arxiv.org/pdf/2409.01610

Cryptocurrency Market Efficiency and Triangular Arbitrage

This research paper examines the efficiency of cryptocurrency markets by analyzing the presence and exploitability of triangular arbitrage opportunities on the Binance exchange, using high-frequency data for Bitcoin, Litecoin, and the U.S. dollar. 
Paper: https://www.sciencedirect.com/science/article/pii/S154461232401537X

TradExpert: Revolutionizing Trading with Mixture of Expert LLMs

"TradExpert," a novel quantitative trading model. TradExpert utilizes a "mixture of experts" approach, employing several specialized large language models (LLMs) to analyze diverse financial data (news, market data, etc.).
Paper: https://arxiv.org/pdf/2411.00782

Deep Learning for Options Trading: An End-To-End Approach

This paper is a discussion about a research that explores the application of deep learning to options trading. The paper proposes a data-driven approach that directly learns from market data.Paper: https://arxiv.org/pdf/2407.21791

The Nonlinear Mind: Innovations in fMRI Analysis and AI Continual Learning

The two papers discussed while from different fields, share a common thread in their approach to handling complex data:
Both studies highlight the limitations of relying solely on linear models or simple replay mechanisms.
fMRI Paper: https://arxiv.org/pdf/1207.3520
Self-recovery of memory via generative replay: https://arxiv.org/pdf/2301.06030

Artificial Intelligence-Aided Protein Engineering (a more accessible discussion w/ Alex & Riley)

In our previous episode, Dr. Ellis and Dr. Chen provided a comprehensive overview of AI-powered protein engineering and Topological Data Analysis. Now, Alex and Riley are here to break down those complex concepts into a more digestible discussion.
 
Paper: https://arxiv.org/pdf/2307.14587

Artificial Intelligence-Aided Protein Engineering: From Topological Data Analysis to Deep Protein Language Models (w/ Dr. Ellis and Dr. Chen)

In this episode, Dr. Ellis and Dr. Chen delve into the exciting world of AI-powered protein engineering. We'll explore how cutting-edge techniques, such as Topological Data Analysis (TDA) and Deep Protein Language Models, are revolutionizing our understanding of protein structure and function.
Paper: https://arxiv.org/pdf/2307.14587

Brain Computer Interface Technology for a Future Battlefield

This discussion is on a fascinating paper exploring the intersection of brain-computer interfaces (BCIs) and future warfare. The paper proposes a BCI system for soldiers, enabling direct brain control of unmanned equipment.Paper: https://arxiv.org/pdf/2312.07818