expressivity

A summary of my activities during the past four years

Entanglement characterization of variational quantum circuits using a Matrix Product State simulator and qiskit.

Scalable Expressiveness through Preprocessed Graph Perturbations (CIKM 2024)

TCR embedding expressivity evaluation framework.

Nearest Neighbor at Every Length: Rank-One Softmax and Hardmax Transformers DOI: https://doi.org/10.5281/zenodo.21698621

This is the repository for a Dash app that helps visualize how layer weights change the surface of the neural network, as well as the arc length of a trajectory

Exploring the expressiveness of higher-order neural networks (activation-less, and spectrally parameterized). This work has been presented at COMPENG 2026

minGRU (Feng et al.) in PyTorch, plus a measured ladder of state-tracking recoveries — signed, rotation, Givens, and delta (DeltaNet-style Householder products) mixers with time decay, all under one parallel scan — with fused Triton GPU kernels as an optional backend