Portrait
Chashi Mahiul Islam
Ph.D. Candidate in Computer Science
Florida State University
About Me

I am a Ph.D. Candidate in Computer Science at Florida State University, advised by Dr. Xiuwen Liu. My research focuses on the robustness and explainability of large multimodal models — studying how microscopic input perturbations cascade through neural representations to cause failures in vision-language understanding, embodied navigation, and reasoning.

I design mechanisms to detect and mitigate these vulnerabilities in Vision Transformers (ViTs), vision-language models (VLMs), and large language models (LLMs). My work has been published at IROS 2024, ISVC 2025, and WCCI-IJCNN 2026, and in Information Sciences (2026). Our paper Neural Expansion: A Unified Mechanism for How Deep Neural Networks Generalize has been accepted at NeurIPS 2026.

In Summer 2026, I interned at Chewy as a Machine Learning Engineer Intern II, working on LLM personalization and evaluation for an AI shopping assistant. In Summer 2025, I interned at Meta, working on vision features for large-scale feed ranking. I also hold an M.Sc. in Computer Science from FSU (GPA: 3.97/4.00) and a B.Sc. in CSE from Bangladesh University of Engineering and Technology (BUET).

Education
  • Florida State University
    Advisor: Dr. Xiuwen Liu · Robustness of Large Multimodal Models
    Ph.D. Candidate in Computer Science
    Jan. 2021 – Dec. 2026 (Expected)
  • Florida State University
    M.Sc. in Computer Science (GPA 3.97/4.00)
    Jan. 2021 – Aug. 2025
  • Bangladesh University of Engineering and Technology
    B.Sc. in Computer Science and Engineering
    2014 – Apr. 2019
Experience
  • Chewy
    Bellevue, WA — Context-aware LLM personalization for an AI shopping assistant
    Machine Learning Engineer Intern II
    Jun. 2026 – Aug. 2026
  • Meta
    Seattle, WA — Vision features and prompt generation for large-scale feed ranking
    Software Engineer Intern
    May 2025 – Aug. 2025
  • Florida State University
    Dept. of Computer Science
    Graduate Research Assistant
    Aug. 2024 – Present
  • Florida State University
    Dept. of Computer Science
    Graduate Teaching Assistant
    Jan. 2021 – Jul. 2024
  • CoKreates Limited
    Dhaka, Bangladesh — ERP system
    Software Developer
    Aug. 2019 – Dec. 2020
Honors & Awards
  • Dean's Award for Doctoral Excellence, College of Arts and Sciences, FSU
    2025, 2026
  • Naaman Franklin Faile Jr. Graduate Fellowship
    2025, 2026
  • GSRC and COGS Travel Grants, FSU — ISVC, Las Vegas
    2025
  • COGS Travel Grant, FSU — IROS, Abu Dhabi
    2024
  • 1st Runner-up, ACM at FSU Fall Programming Contest
    2024
  • 2nd Runner-up, ACM at FSU Fall Programming Contest
    2023
  • Cultural Secretary, Bangladesh Student Association – FSU
    2021
  • President, BUET Film Society
    2018
News
2026
  • Our paper Neural Expansion: A Unified Mechanism for How Deep Neural Networks Generalize has been accepted at NeurIPS 2026.
    2026
  • Interned at Chewy as a Machine Learning Engineer Intern II in Summer 2026, building context-aware LLM personalization for an AI shopping assistant and improving personalization scores by 20% across four rubrics.
    Summer 2026
  • Paper accepted at IJCNN 2026 (IEEE WCCI) (Oral): "Numerical Instability and Chaos: Quantifying the Unpredictability of Large Language Models."
    Jan 15
  • Awarded a NeurIPS 2026 Financial Assistance Award.
    2026
  • 2026
  • Awarded the Naaman Franklin Faile Jr. Graduate Fellowship at Florida State University again in 2026, following the award in 2025.
    2026
  • Honored with the Dean's Award for Doctoral Excellence from the College of Arts and Sciences at Florida State University again in 2026, following the award in 2025.
    2026
2025
  • Completed M.Sc. in Computer Science at Florida State University with a GPA of 3.97/4.00.
    Aug 01
  • Honored with the Dean's Award for Doctoral Excellence from the College of Arts and Sciences at Florida State University.
    May 10
  • Awarded the Naaman Franklin Faile Jr. Graduate Fellowship (2025) at Florida State University.
    May 05
  • Interned at Meta (Seattle, WA) as a Software Engineer Intern (Summer 2025), working on large-scale feed ranking systems.
    May 01
  • Three papers accepted at ISVC 2025: NeuroShield-ViT (Oral), Spatial-ViLT (Oral), and Vision Embeddings & Hallucination Vulnerabilities (Poster).
    Jan 15

All news

Selected Publications (view all )
Neural Expansion: A Unified Mechanism for How Deep Neural Networks Generalize

Chashi Mahiul Islam, Samuel Jacob Chacko, Mao Nishino, Canlin Zhang, Xiuwen Liu

Conference on Neural Information Processing Systems (NeurIPS 2026) · Main Track 2026 Accepted

We proposed Neural Expansion, a geometric account of generalization based on growing directions along which predictions remain unchanged. Growing regions of stable predictions offer a common geometric explanation for several learning behaviors; the mechanism is supported by empirical evidence and partial theory.

Neural Expansion: A Unified Mechanism for How Deep Neural Networks Generalize

Chashi Mahiul Islam, Samuel Jacob Chacko, Mao Nishino, Canlin Zhang, Xiuwen Liu

Conference on Neural Information Processing Systems (NeurIPS 2026) · Main Track 2026 Accepted

We proposed Neural Expansion, a geometric account of generalization based on growing directions along which predictions remain unchanged. Growing regions of stable predictions offer a common geometric explanation for several learning behaviors; the mechanism is supported by empirical evidence and partial theory.

Numerical Instability and Chaos: Quantifying the Unpredictability of Large Language Models

Chashi Mahiul Islam, Alan Villarreal, Mao Nishino, Shaeke Salman, Xiuwen Liu

IEEE World Congress on Computational Intelligence (WCCI 2026), IJCNN 2026 Oral

We discover that microscopic floating-point input perturbations cause up to 10-million-fold amplification through LLM layers, collapsing logit margins and triggering unpredictable token flips — revealing a fundamental numerical chaos regime in large language models. We quantify regimes of stability and rounding-induced chaos to characterize when LLM predictions remain stable or become unpredictable.

Numerical Instability and Chaos: Quantifying the Unpredictability of Large Language Models

Chashi Mahiul Islam, Alan Villarreal, Mao Nishino, Shaeke Salman, Xiuwen Liu

IEEE World Congress on Computational Intelligence (WCCI 2026), IJCNN 2026 Oral

We discover that microscopic floating-point input perturbations cause up to 10-million-fold amplification through LLM layers, collapsing logit margins and triggering unpredictable token flips — revealing a fundamental numerical chaos regime in large language models. We quantify regimes of stability and rounding-induced chaos to characterize when LLM predictions remain stable or become unpredictable.

NeuroShield-ViT: Mechanistic Understandings of Representation Vulnerabilities and Engineering Robust Vision Transformers

Chashi Mahiul Islam, Samuel Jacob Chacko, Mao Nishino, Xiuwen Liu

International Symposium on Visual Computing (ISVC 2025) 2025 Oral

We examine how small input changes affect layer-wise representations in Vision Transformers (ViTs) and propose NeuroShield-ViT, which neutralizes vulnerable neurons without retraining. PGD-100 accuracy improves from 15.6% to 72.3%; classification accuracy also improves by 71.6% against IGO attacks on ImageNet1K.

NeuroShield-ViT: Mechanistic Understandings of Representation Vulnerabilities and Engineering Robust Vision Transformers

Chashi Mahiul Islam, Samuel Jacob Chacko, Mao Nishino, Xiuwen Liu

International Symposium on Visual Computing (ISVC 2025) 2025 Oral

We examine how small input changes affect layer-wise representations in Vision Transformers (ViTs) and propose NeuroShield-ViT, which neutralizes vulnerable neurons without retraining. PGD-100 accuracy improves from 15.6% to 72.3%; classification accuracy also improves by 71.6% against IGO attacks on ImageNet1K.

Malicious Path Manipulations via Exploitation of Representation Vulnerabilities of Vision-Language Navigation Systems

Chashi Mahiul Islam*, Shaeke Salman*, Montasir Shams, Xiuwen Liu, Piyush Kumar (* equal contribution)

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024) 2024 Oral

We develop adversarial algorithms that exploit semantic representation gaps in vision-language navigation models to redirect robotic agents along malicious paths, achieving 91.7% landmark-conditioned path redirection and proposing a noise-sensitivity-based detection mechanism that improves navigation policy stability by 96.3%.

Malicious Path Manipulations via Exploitation of Representation Vulnerabilities of Vision-Language Navigation Systems

Chashi Mahiul Islam*, Shaeke Salman*, Montasir Shams, Xiuwen Liu, Piyush Kumar (* equal contribution)

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024) 2024 Oral

We develop adversarial algorithms that exploit semantic representation gaps in vision-language navigation models to redirect robotic agents along malicious paths, achieving 91.7% landmark-conditioned path redirection and proposing a noise-sensitivity-based detection mechanism that improves navigation policy stability by 96.3%.

All publications