Artificial intelligence research portfolio

Sakib Hossain

Email Google Scholar GitHub LinkedIn CV

Researcher · builder · mentor

Generative and visual AI

3D vision and world models

About

I am an M.S. student in Artificial Intelligence at KAIST and a researcher at the Statistical Artificial Intelligence Lab (SAIL), Kim Jaechul Graduate School of AI, South Korea, advised by Professor Jaesik Choi.

Before joining KAIST, I received my B.Sc. in Computer Science and Engineering from American International University-Bangladesh (AIUB). My research experience has also included work with The University of Queensland, Australia and Qatar University, Qatar.

I am also the Co-Founder & CEO of NeuronTreeAI, a nonprofit initiative co-founded with Prof. Mohammad Ali Moni focused on AI education, research mentorship, and making advanced research knowledge more accessible.

My long-term research goal is to become a professor at a top university and launch an AI startup that empowers society for a better future. My primary immediate goal is to secure a PhD position for the Fall 2027 intake at a top-10 university in the United States.

For research discussions and collaborations, contact sakib@kaist.ac.kr.

Research works

Selected publications

All publications

Orthogonal Polynomial Approximation for Matrix Log Normalization in Global Covariance Pooling

Md Rifat Ur Rahman, Md Raihan Khan, Md Sakib Hossain Shovon, Pietro Liò, and Mohammad Ali Moni

In British Machine Vision Conference (BMVC), 2026

Accepted at BMVC 2026 · Accepted August 21, 2026

Abs Bib

@inproceedings{rahman2026orthogonal,
  title = {Orthogonal Polynomial Approximation for Matrix Log Normalization in Global Covariance Pooling},
  author = {Rahman, Md Rifat Ur and Khan, Md Raihan and Hossain Shovon, Md Sakib and Li{\`o}, Pietro and Moni, Mohammad Ali},
  booktitle = {British Machine Vision Conference (BMVC)},
  year = {2026},
  note = {Accepted at BMVC 2026 · Accepted August 21, 2026},
}

From Image to Language: A Critical Analysis of Visual Question Answering Approaches, Challenges, and Opportunities

M. F. Ishmam, Md Sakib Hossain Shovon, M. F. Mridha, and N. Dey

Information Fusion, 2024

Published January 28, 2024

Abs Bib

A critical study of visual question answering architectures, multimodal reasoning strategies, limitations, and future research opportunities.
@article{ishmam2024image, title = {From Image to Language: A Critical Analysis of Visual Question Answering Approaches, Challenges, and Opportunities}, author = {Ishmam, M. F. and Hossain Shovon, Md Sakib and Mridha, M. F. and Dey, N.}, journal = {Information Fusion}, year = {2024}, note = {Published January 28, 2024}, }

Conducted projects

Research projects

Current and prior applied research initiatives.

SDM Manufacturing Foundation Model Project

An ongoing, large-scale manufacturing foundation model initiative involving leading Korean research institutions.

Project

Automatic Bangla Number Plate Recognition for Smart Toll Collection

An AI project for automatic Bangla vehicle number plate recognition in intelligent toll collection infrastructure.

Project

View all projects

Research at a glance

Core interests

Diffusion & Flow Matching

Generative modeling, stochastic interpolants, sampling, scheduler optimization, and efficient generation.

Physical AI & World Models

Visual world modeling, physical reasoning, embodied and interactive AI, and scene dynamics.

3D Vision

3D representation learning, geometry-aware learning, scene understanding, reconstruction, and rendering.

Explainable AI

Mechanistic interpretability, neuron polysemanticity, representation analysis, and concept disentanglement.

Inference Optimization

Efficient inference, few-step generation, model acceleration, and test-time optimization.

Multimodal Foundation Models

Vision-language models, multimodal reasoning, cross-modal representations, and foundation model adaptation.

Updates

All news

2026	Submitted my M.S. thesis work to ICLR 2027 from SAIL, KAIST AI.
2026	Submitted a research work to WACV 2027 in collaboration with SAILab (KAIST AI) and NeuronTreeAI.
2026	Submitted a research work to AAAI 2027 in collaboration with SAILab (KAIST AI), KAIST Visual AI Group, and NeuronTreeAI.