Stevens Institute of Technology

ECAI Lab

Explainable and Controllable AI

Welcome! We are the Explainable and Controllable AI Lab at Stevens Institute of Technology. We study how AI works, how to control its behavior, and how the understanding can make AIs fast, safe and reliable.

Our work connects mechanistic interpretability with reasoning, efficient computation, and AI agents for research and education.

01 / Research

Research pillars, from insight to impact

01

Understand the mechanisms

Probe language models to connect their internal mechanisms with behavior and safety.

02

Guide the behavior

Develop behavioral and mechanistic methods for controlling AI systems.

03

Reasoning AI

Build truthful AI that helps researchers and students discover and learn.

04

Bring agents into the world

Explore AI agents for finance, cybersecurity, education, and embodied applications.

Research in practice

HomeBot One

We collaborate with HomeBot One on efficient vision-language-action models for home robots.

Scholera Inc.

Our research has translated into Scholera Inc., bringing AI into learning and education.

With support from
From the lab

Latest news

Awards, milestones & updates
We receive a Tinker Research Grant. Thanks Thinking Machines Lab!
We receive an NVIDIA Academic Grant Program award. Thanks Nvidia!
Our project Scholera wins an New Jersey PlugNPlay AI Innovation Challenge award, and is spinning towards a startup. Congrats Kevin and Dr. Hao Wang!
Shashidhar receives a Nokia Bell Labs Scholarship.
We receive an NSF NAIRR Pilot award for our AI4Research project.
We receive an NSF NAIRR Pilot award for our Explainable AI Acceleration project.
We receive a Google Cloud Research Credits award. Thanks Google!
Our paper "ACCORD: Closing the Commonsense Measurability Gap" receives an Outstanding Paper Award at NAACL.
We receive a Stevens Teaching Learning Center award for the AI4Education project.
A new course, Topics in Explainable Natural Language Processing, is developed for the Spring 2025 semester.
Selected work

Selected papers

Full list at Google Scholar

2026Safety

Towards Fine-grained and Robust Detection of Human–AI Co-authorship

Findings of EMNLP 2026

Paper link forthcoming

5 selected publications

Our people

Team

Current members 22

Mack Arif

Mack Arif

Masters · CompSci @ Stevens

Olin d'Souza

Olin d'Souza

Masters · CompSci @ Stevens

Husain Aliasger Dhansura

Husain Aliasger Dhansura

Masters · CompSci @ Stevens

Ynez Prem Dias

Ynez Prem Dias

Masters · CompSci @ Stevens

Umer Rizwan

Umer Rizwan

Undergrad · CompSci @ Stevens

Alex Santora

Alex Santora

Undergrad · CompSci @ Stevens

Asher Egerton-Idehen

Asher Egerton-Idehen

Undergrad · CompSci @ Stevens

Alumni 32

Muhammad Shahzaib Chaudhary

Muhammad Shahzaib Chaudhary

Masters · CompSci @ Stevens

Rohit Sandadi

Rohit Sandadi

Dougherty Valley High School

Surya Gedipudi

Surya Gedipudi

High Technology High School

Sam Pan

Sam Pan

Undergrad · CompSci @ Washington St Louis

Daniel Wan

Daniel Wan

Undergrad · CompSci @ NEU

Jim Yang

Jim Yang

Undergrad · EngSci @ Toronto

Zining Zhu (朱子宁)

Assistant Professor
School of Computing
Gateway South 429

[email protected]

Zining Zhu is an Assistant Professor in the School of Computing at Stevens Institute of Technology, where he directs the Explainable and Controllable AI Lab. He received his Ph.D. from the University of Toronto and Vector Institute, advised by Frank Rudzicz. He is affiliated with SIAI and CRAFT at Stevens. His research spans natural language processing and explainable AI, connecting an understanding of model mechanisms with safer, more reliable, and efficient AI systems. He has received a paper award at NAACL and served as Senior Area Chair for EMNLP, Senior Program Committee for AAAI, Area Chair for NeurIPS, ICML, and ICLR, and an Action Editor for ACL Rolling Review.