Shifat E. Arman

Shifat E. Arman

Assistant Professor

Department of Robotics and Mechatronics Engineering, University of Dhaka

Founding Director

Cortex AI Research Group

I am an Assistant Professor in Robotics and Mechatronics Engineering specializing in developing artificial intelligence for real-world applications. I have authored 14 peer-reviewed articles, including 8 in Q1 venues, with over 355 citations, and have received multiple research awards and four competitive research grants totaling over 2 million BDT. I also founded the Cortex AI Research Group, which conducts cross-disciplinary work in medical imaging, agriculture and robotics. My research interests include Agentic AI, Multimodal and Multitask Learning, Vision Language Models, and Bio-Inspired Robotics, with an emphasis on interpretable AI for real-world applications.

Google Scholar
View profile →
355
Citations
8
h-index
7
i10-index

Life Events

2026Paper accepted to ACL 2026 (Findings)Thinking Like a Botanist: Challenging Multimodal Language Models with Intent-Driven Chain-of-Inquiry. ACL is the #1 ranked venue in Computational Linguistics (h5-index 236, CORE A*).
2026Decipede, our bio-inspired modular myriapod robot, won a Silver Medal at the World Invention Creativity Expo (WICE) in Malaysia and placed 4th at the Global Robotics and Innovation Challenge (GRIC) in Turkey.
2025TurtleBot, our sea-turtle-inspired underwater robot, became Champion of Robotronics Fest 2025.
2025Received the Dean's Award (2025 and 2023) as one of the youngest awardees and only lecturer to receive it twice.
2025Promoted to Assistant Professor, Dept. of Robotics and Mechatronics Engineering, University of Dhaka.
2025Received the Provost Award for securing 1st position in the MSc with a CGPA of 3.98 out of 4.00.
2025Received the Dhaka University Research Excellence Award as one of the youngest awardees for impactful peer reviewed research.
2024Received the Special Merit Award and Distinction for securing 2nd position among 58 faculty members in the Foundation Training Programme.
2022Joined as Lecturer, Dept. of Robotics and Mechatronics Engineering, University of Dhaka.
2022Joined as Adjunct Faculty, Dept. of Computer Science and Engineering, BRAC University.
2022Completed MSc in Robotics & Mechatronics Engineering from University of Dhaka.
2021Joined as Visiting Research Assistant, GIS and Remote Sensing Laboratory, Gazipur Agricultural University.
2020Joined as Research Assistant, University of Dhaka.
2019Completed BSc in Robotics & Mechatronics Engineering from University of Dhaka.
2019Joined as AI Design and Development Intern, Datasoft Manufacturing and Assembly Limited.

Featured Works

Full list of published articles on Google Scholar →

Decipede — Bio-Inspired Modular Myriapod Robot
🥈 Silver Medal · WICE4th Place · GRIC

Decipede (Flexipede)

Bio-inspired Modular Myriapod Robot for Rough-Terrain Traversal

A modular robot inspired by centipede and millipede locomotion, designed for adaptive traversal of rough and unstructured terrain. Each segment responds independently to obstacles, enabling coordinated undulatory gait patterns. In 2025, Decipede earned two international honors: a Silver Medal at the World Invention Creativity Expo (WICE) in Malaysia and 4th place at the Global Robotics and Innovation Challenge (GRIC) in Turkey.

TurtleBot — Sea Turtle-Inspired Underwater Robot
🏆 Champion · Robotronics Fest 2025

TurtleBot

Sea Turtle-Inspired Underwater Exploration Robot

An underwater exploration robot designed around the biomechanics of sea turtle locomotion, using bio-inspired flipper actuation for efficient underwater navigation. Won the Champion title at Robotronics Fest 2025. Funded by the ICT Division Innovation Fund, Bangladesh, as part of the smart biomimetic robots for underwater operations research grant.

PlantInquiryVQA · Chain-of-Inquiry FrameworkSakib et al., ACL 2026 FindingsPhase 1 · Visual Cue ExtractionSunflower · Alt. Leaf SpotQwen3-VL-4BExpert SchemaGrounded Visual Cueschlorosis (mild)lesion margin: definedconcentric ringsyellow halo presentnecrosis: 15-30%GPT-4V cross-validation · 93.8% factualityPhase 2 · Intent DeterminationDisease Severityearlyintermediateadvanced↓ maps to epistemic intent k ∈ Kk_D · DiagnosisAmbiguity resolution · Differential dxk_P · PrognosisFuture forecasting · Etiologyk_M · ManagementAction planning · CounterfactualPhase 3 · CoI GenerationLLM · Reasoning Modules12 templates · 7 categoriesC(x, vₓ) =q₁: morphology?a₁: oval, serratedq₂: lesion type?a₂: concentric ringsq₃: prognosis?a₃: advancingconditioned on vₓ, Hₜ₋₁, kFormalizationC(x, vₓ) = ⟨(q₁, a₁), (q₂, a₂), … , (q_T, a_T)⟩where each qₜ is conditioned on visual evidence vₓ, history Hₜ₋₁, and latent diagnostic intent k ∈ { k_D, k_P, k_M }PlantInquiryVQA Benchmark · Key Numbers24,950expert-curatedplant images138,068question-answerpairs7cognitivecategories12reasoningtemplates93.8%factuality onclinical eval.3epistemicintentsVisual cues → epistemic intent → adaptive question trajectorybenchmarking MLLMs on multi-step, evidence-grounded botanical reasoningFindings: structured inquiry reduces hallucination and improves diagnostic correctness
ACL 2026 (Findings)Multimodal LMsVQA Benchmark

Thinking Like a Botanist

Challenging Multimodal Language Models with Intent-Driven Chain-of-Inquiry

We introduce PlantInquiryVQA, a benchmark of 24,950 expert-curated plant images and 138,068 question-answer pairs for multi-step, intent-driven visual reasoning in botanical diagnosis. We formalize a Chain-of-Inquiry (CoI) framework that models expert diagnostic trajectories as ordered question-answer sequences conditioned on grounded visual cues and three epistemic intents — diagnosis, prognosis, and management. Evaluations on top-tier MLLMs show that structured, question-guided inquiry significantly improves diagnostic correctness and reduces hallucination. Accepted at ACL 2026 (Findings).

SugarcaneShuffleNet

SugarcaneShuffleNet

AI for Sugarcane Leaf Disease Diagnosis

The SugarcaneAI app for diagnosis of 15 prominent sugarcane leaf diseases using SugarcaneShuffleNet, a lightweight CNN optimized for mobile deployment in field conditions.

Intracranial Hemorrhage

Intracranial Hemorrhage Classification

Deep Learning and Bayesian Optimization on CT Scans

Intracranial Hemorrhage classification from CT scans using deep learning and Bayesian Optimization for robust, hyperparameter-tuned medical image analysis.

Research Portfolio

Artificial Intelligence in Healthcare

  1. 1.AI for COVID-19 Diagnosis from X-ray Images
  2. 2.AI for Intracranial Hemorrhage Diagnosis from CT Scans
  3. 3.AI for Histological Cancer Cell Detection from Tissue Image
  4. 4.AI for Retinal Disease Diagnosis from OCT Images

Artificial Intelligence in Agriculture

  1. 1.AI for Sugarcane Leaf Disease Diagnosis from Mobile Images
  2. 2.AI for Banana Leaf Disease Diagnosis from Mobile Images (BananaSqueezeNet)
  3. 3.AI for Mango Leaf Disease Diagnosis from Mobile Images
  4. 4.AI for Diverse Crop Leaf Diagnosis from Mobile Images (PlantCareNet)
  5. 5.AI for Cotton Seedling Monitoring and Growth Stage Classification

Artificial Intelligence in Nanophotonics

  1. 1.AI for Inverse Design in Nanophotonics
  2. 2.Generative AI for Inverse Design in Nanophotonics

Research Grants

2025–26·Associate Investigator

Deep Learning and XAI for Interpretable Microscopic Tissue Image Analysis for Cancer Diagnosis

Ministry of Science and Technology, Bangladesh

2024–25·Principal Investigator

Building Smart Biomimetic Robots for Underwater Operations

ICT Innovation Fund, ICT Division, Bangladesh

2023–24·Co-Investigator

Rapid Assessment of Plant Responses to Climate-Induced Stress Using AI

University Grants Commission of Bangladesh

2023–24·Co-Investigator

SmartPDE: Deep Learning Based Plant Disease Epidemiology and Forecasting

University Grants Commission of Bangladesh

Awards

Dhaka University Research Excellence Award 2025

2025
  • One of the youngest awardees, recognised for impactful peer-reviewed research contributions.

Dean's Award 2023 and 2025

2023, 2025
  • One of the youngest awardees and only lecturer to win twice, recognising research, mentorship and co-curricular contribution.

Provost Award 2025

2025
  • Received the Provost Award for securing first position in the MSc programme with a CGPA of 3.98 and the highest marks.

Special Merit Award and Distinction

2024
  • Received the award for securing the second highest score among 58 faculty members in the 14 day Foundation in University Teaching and Learning training and was one of the youngest participants.

Education

MSc, Robotics & Mechatronics Engineering

University of Dhaka

Thesis: Algorithms for Fast and Efficient Neural Architecture Search

First position · CGPA 3.98 · Provost Award (2025)

BSc, Robotics & Mechatronics Engineering

University of Dhaka

Thesis: Deep Learning in Medical Image Analysis

Teaching

Undergraduate and graduate courses.

Artificial Intelligence

  1. Artificial Intelligence
  2. Machine Learning
  3. Neural Networks and Deep Learning

Robotics & Programming

  1. Human–Robot Interaction
  2. Introduction to Robotics
  3. Fundamentals of Programming
  4. Object Oriented Programming

Mechatronics

  1. Advanced Mechatronics Engineering
  2. Programmable Logic Controller
  3. Engineering Drawing

Collaborations

  • Ecology Lab, Dept. of Botany, University of Dhaka
  • GIS & Remote Sensing Lab, Gazipur Agricultural University
  • Microsystems & Nanoengineering Lab, Dept. of Electrical and Electronics Engineering, University of Dhaka
  • Dept. of Geology, University of Dhaka

Contact

Department of Robotics and Mechatronics Engineering, University of Dhaka, Dhaka 1000, Bangladesh