Featured Projects

• Recall Distortion in Neural Network Pruning and the Undecayed Pruning Algorithm
  Paper accepted into the NeurIPS 2022 Conference.
  In this work, we studied relative distortions in class-level recall caused by network pruning by hypothesizing an intensification effect.
  In layman's terms: we studied how removing weights from ML networks can cause very unequal class-level accuracies even when the model-level accuracy remains high.
• McCoy 8-bit Microprocessor
  Created a tiny RISC processor with 7 instructions, 10 registers, and can run a 256 instruction long program. Taped-out and submitted for fabrication on a chipIgnite shuttle through the Tiny Tapeout program.
• GitHub Profile
  Whatever makes it onto there