My adventures in pushing the boundaries of technology through experiments.
Research Paper • 2024
Standard self-attention in transformers scales quadratically with sequence length, limiting their application to long documents. We survey recent approaches to achieve linear or ne
Research Paper • 2024
Federated Learning (FL) enables training machine learning models on distributed data without centralizing it. While this provides inherent privacy benefits, additional techniques a
Research Paper • 2024
Effective exploration remains a fundamental challenge in reinforcement learning. We review classical and modern exploration strategies, analyzing their theoretical properties and e