Center for Applied Artificial Intelligence

Where beginners
become researchers.

A guided research lab working across multimodal AI, vision–language models, generative systems and the ethics that hold them accountable — built so that first-time researchers can publish real work.

Operating under The Korean Academy Video is property of Northwestern University
Scroll

01 — The Lab

Founded
2022
Operating under
The Korean Academy
Registered in
India

CAAI is an experimental research lab for students entering artificial intelligence, machine learning, computer vision and cross-modal systems.

We provide a guided environment where beginners learn research methodology, implementation practice, paper writing and experimental design under structured mentorship — not by watching, but by running projects end to end. Members join a cohort, take a question, and carry it through to a preprint, a journal submission or a released dataset.

The lab supports students learning Korean and preparing for academic or professional transitions to Korea and elsewhere, while building research profiles for international higher education.

Read about the lab →

54
Researchers & alumni
12
Research projects
12
Papers & datasets
04
Research pillars

02 — Research Areas

Four pillars, one question: how machines make meaning.

01

Multimodal AI

Merging vision, text and audio into systems that read the world the way people do — visual question answering, captioning, and cross-modal alignment.

02

AI Ethics & Bias

Fairness, accountability and transparency in machine learning — measuring where models fail people, and designing evaluations that catch it early.

03

Generative Models

Large language models and diffusion systems for generating text, images, audio and code, and the architectures that make them controllable.

04

AI for Social Good

Applying the same techniques to healthcare, education and environmental problems where better models translate into better outcomes.

03 — Current Work

Four projects are running in the lab right now.

Image Music
Fig. 1 — Cross-modal encode → generate
In progress

Multimodal Image-to-Music Generation

Learning a mapping from visual input to musical output — encoding an image into a shared representation and decoding it into audio that carries the scene's structure, mood and rhythm.

10 August 2026 — Ongoing Cross-modal generation
Researchers

Gurpreet Singh · Lamia Qamar

Script Panels
Fig. 2 — Narrative text → sequential panels
In progress

Text-to-Comic Generation

Turning written narrative into sequential visual storytelling — segmenting a script into beats, then generating panel layout, imagery and dialogue that stay consistent across the sequence.

Ongoing Multimodal generation
Researchers

Trina Banerjee · Mukhthikka · Gurpreet Singh

Sequence Imaging Target edit
Fig. 3 — Sequence + imaging → targeted edit
In progress

Multimodal AI in Gene Editing

Bringing the lab's cross-modal methods to molecular biology — combining sequence data, imaging and published literature in a single representation to support target selection and outcome prediction in gene editing.

Ongoing Multimodal · Biomedical
Researchers

Sudipta Patil · Gurpreet Singh

Partners Shared work
Fig. 4 — Distributed sites, one research programme
In progress

International Research Collaboration

A collaboration run across borders, pairing the lab's AI methods with research in international business, ESG and media — and testing how a distributed team actually sustains a shared research programme.

Ongoing Cross-disciplinary
Researchers

Ritika Kanojia · Alimpia Roy · Gurpreet Singh

04 — Dispatch

Recent from the lab.

News & Events

Team members attended The European AI Conference 2025.
Team attended SECON International — 4th Annual International ISC2 Chapter Conference.
Gurpreet Singh attended the International Conference on Platform and AI Society in Asia.
Lamia Qamar presented “Comparative Study of Convolutional Neural Networks for Deepfake Video Detection on Social Media” at ICA2S 2025.

05 — Join Us

Open positions, Fall 2025.

CAAI is accepting graduate research assistants and undergraduate research interns. We look for strong programming ability — Python, PyTorch or TensorFlow — and the patience research actually takes.