Ran Zhang

Counterbalancing Teacher: Regularizing Batch Normalized Models for Robustness

2022 · arXiv (Cornell University)

Batch normalization (BN) is a ubiquitous technique for training deep neural networks that accelerates their convergence to reach higher accuracy. However, we demonstrate that BN comes with a fundamental drawback: it incentivizes the model to …

Applying NLP Techniques to Classify Businesses by their International Standard Industrial Classification (ISIC) Code

2022 · 2022 IEEE International Conference on Big Data (Big Data)

The application of machine learning has played an important role in several aspects of text classification across domains, and has brought with it great changes to the current state of the art. In this paper, …

Spatial Target Tracking Algorithm Based on Heterogeneous Sensors Data Fusion

2023

The track of Spatial targets is basically non-linear, which makes the cost of using homogeneous sensors for space target tracking significantly increase. Data fusion using heterogeneous sensors can achieve information complementarity, thereby improving recognition ability. …

Deep Cut-informed Graph Embedding and Clustering

2025 · arXiv (Cornell University)

Graph clustering aims to divide the graph into different clusters. The recently emerging deep graph clustering approaches are largely built on graph neural networks (GNN). However, GNN is designed for general graph encoding and there …

Probabilistic greedy algorithm solver using magnetic tunneling junctions for traveling salesman problem

2026 · Open MIND

Combinatorial optimization underpins applications in artificial intelligence, logistics, and network design, yet classical techniques such as greedy search and dynamic programming struggle to balance efficiency and solution quality at scale. We present a probabilistic framework …