Hopkins Brains Bots and Behavior Lab

Johns Hopkins campus with spring blossoms

Brains Bots and Behavior (B3) Lab

Learning from human data for embodied intelligence.

We bring together disciplines across observational and interactive learning to develop embodied systems that can reason about people, learn from their interactions, and operate around them.

Brains
Bots
Behavior

Research

Human behavior as a foundation for intelligent action.

Dexterous robot with articulated hands observing a person preparing food in a lab kitchen

Watch

Learning From Human Data

Models that learn from demonstrations, video, motion, gaze, language, and interaction traces to capture the what, how, and why of human interactions.

Hands and objects tracked in a behavior understanding study

Understand

Behavior Understanding

Representations for goals, intent, preferences, affordances, and context in everyday behavior, for enabling multimodal predictions.

Smart glasses view with a sequence of translucent future hand-pose predictions reaching toward an object

Predict

Embodied Prediction

Forecasting human motion, object changes, scene variations through structured predictive models so that AI systems know how and when to assist humans.

Robot handing a small object to a person

Act

Robots Around People

Closing the loop from learned predictive models to physical systems capable of dexterous and long-horizon mobile manipulation while acting with awareness of people, tasks, uncertainty, and shared spaces.

People

Homanga Bharadhwaj

Principal Investigator

Homanga Bharadhwaj

Assistant Professor of Computer Science
Johns Hopkins University

PhD Students

Carter Ung

1st Year CS PhD

Carter Ung

Yash Jangir

1st Year CS PhD

Yash Jangir

Michael (Suyu) Ye

1st Year CS PhD

Michael (Suyu) Ye

Co-advised by Tianmin Shu.

BS / MS Students

Yuqi Tang

Daniel Guo

Daniel Guo

PhD Student Collaborators

Videos

Video highlights of past research from the PI connecting human video, prediction, and robot action.

ECCV 2024

Track2Act

Point-track prediction from internet videos for zero-shot robot manipulation.

ECCV 2026

ObjectForesight

Future 3D object trajectory prediction from egocentric human videos.

arXiv 2026

MotionForesight

Future 3D scene-flow prediction from short monocular human-interaction videos.

TMLR 2025

HandsOnVLM

Predicting future interaction trajectories of human hands in a scene given high-level colloquial task specifications in the form of natural language.

ICRA 2026

AINA

Learning multi-fingered robot manipulation policies directly by watching videos of humans with Aria glasses on, without any robot interaction/tele-operation/simulation data.

IROS 2026

SPIDER

A physics-based retargeting framework to transform and augment kinematic-only human demonstrations to dynamically feasible robot trajectories at scale.

Publications

News

B3 Lab website is launched.

Contact Us

Thank you for your interest in our lab.

For research-related inquiries, please reach out to Prof. Homanga Bharadhwaj.

Joining B3

Information for prospective lab members and visitors

01

Prospective PhD Students

We will consider new PhD students each year. Please apply through the Johns Hopkins Computer Science PhD admissions page and mention your interest in working with Prof. Homanga Bharadhwaj. We are unable to reply to individual admissions emails before decisions are released.

02

Prospective Postdocs

If you are interested in a postdoctoral position, please submit the lab interest form below and also email Prof. Homanga Bharadhwaj with a brief note.

03

JHU Undergraduate or Master’s Students

Current or admitted Johns Hopkins undergraduate and master’s students interested in research, independent study, or thesis opportunities should submit the lab interest form below.

04

Prospective Visitors

If you are not a current Johns Hopkins student and are interested in a short-term research visit, please submit the lab interest form below.

Open the lab interest form

The Google Form link will appear here when it is added to the site configuration.