Md. Shihab Shahriar

Uttara, Dhaka 1230, Bangladesh | +8801854243064

I am an aspiring researcher with a passion for advancing the field of machine learning and natural language processing. Driven by Nelson Mandela's words, "A winner is a dreamer who never gives up," I approach life's challenges with an adaptable spirit and determination to keep going. There are always new skills to build, innovations to uncover, and boundaries to push.

Research Interests

Grounding Language Models

I want to contribute to the development of language models that not only understand linguistic nuances but also effectively connect with real-world contexts. Two key areas I am interested in are mentioned here.

  1. Commonsense Reasoning and Knowledge Integration:

    Exploring methods to seamlessly integrate commonsense reasoning and background knowledge into large language models. Aim to enhance model capabilities for making informed and contextually appropriate responses.

  2. Interactive and Situated Learning:

    Researching strategies to enable language models to learn interactively and adapt to dynamic environments. Such as learning from simulated environments, image, audio, video etc.

Efficient Training of Language Models

As I faced repeated challenges with available memory and compute, I am also deeply interested in advancing the accessibility of language models by reducing computation and memory costs during training.

  1. Parameter-Efficient Fine-Tuning:

    Investigating methods for fine-tuning language models with a focus on efficiency. Also exploring the integration of low-rank adapters and adapter modules to achieve fine-tuning with a reduced number of parameters.

  2. Model Size Reduction Techniques:

    Make models more accessible by reducing the size and complexity of language models. This includes delving into techniques like quantization, pruning and distillation.

Publications

Question Answer Generation in Bengali: Mitigating the scarcity of QA datasets in a low-resource language

AACL-IJCNLP 2023

Md Shihab Shahriar, Ahmad Al Fayad Chowdhury, Md. Amimul Ehsan and Abu Raihan Kamal

Paper Link
2023

HateXplain2.0: An Explainable Hate Speech Detection Framework Utilizing Subjective Projection from Contextual Knowledge Space to Disjoint Concept Space

ICML 2023 Workshop AI&HCI

Md Fahim, Md Shihab Shahriar, Mohammad Sabik Irbaz, Syed Ishtiaque Ahmed, Mohammad Ruhul Amin

Paper Link | Poster Link
2023

Unveiling the Essence of Poetry: Introducing a Comprehensive Dataset and Benchmark for Poem Summarization

EMNLP 2023

Ridwan Mahbub, Ifrad Towhid Khan, Samiha Shafiq Anuva, Md Shihab Shahriar, Md Tahmid Rahman Laskar, Sabbir Ahmed

Paper Link
2023

Experience

Lecturer

Islamic University of Technology

Taught courses in Data Mining and Engineering Ethics, and served as a Lab Instructor for RDBMS, Machine Learning, and Pattern Recognition Labs.

Aug 2022 – Present

Software Engineer

Enosis Solutions

Maintained and updated a software which monitored and orchestrated customer journeys. Learned important tools and techniques such as Angular, Java Spring Framework, Bitbucket and Jira.

Jun 2022 – Aug 2022

Machine Learning Engineer Intern

Pioneer Alpha

Worked on a research project involving Natural Language Inference. One of the experiences that made me explore the NLP domain and ignite my initial curiosity about the domain.

Apr 2021 – Oct 2021

Education

Islamic University of Technology

B.Sc. in Computer Science and Engineering

CGPA: 3.96/4.00 (2nd in Class of 69)

2018 – 2022

Important Test Scores

GRE General Test: Q-166, V-159, AW-4.0 (325/340)

TOEFL iBT Test: 110/120 (R-29, L-29, S-27, W-25)

Skills

Python
Java
C++
HTML
CSS
JavaScript
MySQL
MongoDB
Natural Language Processing
Competitive Programming

Interests

Apart from being a lecturer, I enjoy outdoor activities. I really like playing football, tennis, and badminton. I also like traveling and enjoy exploring places.

When indoors, I like playing table tennis and watching movies. I spend a large amount of my free time exploring the latest advancements in the "Natural Language Processing" world.

Awards & Certifications

  • Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization

    Coursera

  • Neural Networks and Deep Learning

    Coursera

  • Structuring Machine Learning Projects

    Coursera

  • Datathon 2.0 - Robi

    Finalist (Top 21)

  • CSE Festival Week IT OLYMPIAD (ICT Quiz Contest) - City University

    1st Prize

  • Ennovators 3.0 - Banglalink

    Top 40