Research
  • Forensic Analysis of Bangla Handwritten Letters
    Publisher: Proceedings of the 8th International Conference on Computer and Communications Management
    Conference: ICCCM'20

    Abstract: Handwriting says a lot about a person. Sometimes the information hidden there becomes the most important clues for an investigation. Traditionally detectives take help from the Graphologists. However, given that handwriting analysis is mostly based on visual features, machine learning algorithms should be able to find out important features. In this research, we used a Bangla handwriting dataset to identify the age, gender, and location (district) of the writer using deep learning algorithms. To the best of our knowledge, we are the first to address these three features from Bangla handwriting. We found that age could be identified with around 87.2% accuracy, location with 65.2% accuracy, and gender with 55.8% accuracy. The main causes of the low accuracy are the complex geometric shapes of Bangla letters. Mastery of those shapes is clearly reflected by the age of the writers.
    DOI: 10.1145/3411174

  • Soil Analysis and Unconfined Compression Test Study using Data Mining Techniques
    Publisher: ICCCI 2020: Advances in Computational Collective Intelligence
    Conference: 12th International Conference on Computational Collective Intelligence (ICCCI 2020)
    Abstract: In this study, Random Forest Regressor, Linear Regression, Generalized Regression Neural Network (GRNN) and Fully connected Neural Network (FCNN) models are leveraged for predicting unconfined compression coefficient with respect to standard penetration test (N-value), depth and soil type. The study is focused on a particular correlation of undrained shear strength of clay (Cu) with the standard penetration strength. The data used is from 14 no. ward in Mymensingh and Rangamati districts which are situated in Bangladesh. By using this data, the study tries to solidify the correlation of SPT (N-value) with Cu. It also tries to check the goodness of the relationship by comparing it with unconfined compression strength values gained from the unconfined compression test calculated from the field by experts. DOI: 10.1007/978-3-030-63119-2_4

  • Analysis of Soil and Various Geo-technical Properties using Data Mining Techniques
    Publisher: IEEE
    Conference: 2020 IEEE 10th International Conference on Intelligent Systems (IS)

    Abstract: In this study, General Regression Neural Network(GRNN), Artificial Neural Network (ANN), Fully Connected Neural Network (FCNN), Support Vector Regression (SVR) and Linear Regression (LR) models have been implemented in order to predict the composition of soil with respect to the Standard Penetration Test (SPT), and soil depth. The primary focus has been on determining a significant correlation between the soil composition with SPT value and depth. Data sets have been used from ward 14, Mymensingh district of Bangladesh and from a construction project along India-Myanmar border. In this study, 8 types of soil, namely, fine sand, silty clay, clayey silt with fine sand, clayey silt, fine sand with silt, silty fine sand, sandy silt, and rubbish has been classified, and the probability of obtaining the soil type classification has been determined.
    DOI: 10.1109/IS48319.2020.9199941


Courses
  • Junior Design | Course Code: CSE 299
    Project Idea— At present times we have various web based tools like google-classroom and piazza to manage resource with a large group of people. These are specially used by educational institutions (e.g. school, college, university) and other organization to communicate with a large number of people, to create workflow etc. But often using different sites simultaneously causes a lot of clutter and becomes cumbersome to keep track of. Hence we offer ‘Populace’ which is a web based application. The reason it stands out from the rest is because it will combine all this separate existing platform into one single platform. By signing in to ‘Populace’ users will be able to see posts made on the other existing web applications and also make their own query. It will therefore be a gathering place especially for students and for people who wants a one-stop solution to keep track of all the accounts in different web platform. Thus creating a better workflow. Primarily, the two platforms that we will include are ‘Google Classroom’ and ‘Piazza’.
    Click to visit github repo of the project
    Project Members: Sayeed Md. Shaiban and Monisha Saha

  • Natural Language Processing | Course Code: CSE 495
    Abstract—BERT, Bidirectional Encoder Representation from Transformers is a masked language model that has created a stir in the Machine Learning community by exhibiting state of the art results in about a wide assortment of NLP problems, including Question Answering (SQuAD v1.1), Natural Language Inference (MNLI), and others. In this project , we use BERT to extract features and word embeddings from Bangla text input and created a visualization tool to understand the nuances that the BERT model is able to retain form Bangla language
    Click to read the full report
    Project Members: Sayeed Md. Shaiban and Abdullah Md. Sarwar

  • Pattern Recognition | Course Code: CSE 465
    Abstract—Summarization has long been a significant area of concern in the field of Natural Language Processing, mostly due to the its dependency on human intervention. In this study, we use a denoising autoencoder, BART, to carry out abstractive summarization on the dataset of COVID-19, using ROUGE scores as an evaluation metric. The primary focus of the study has been the use of medical articles based on COVID-19, which is aimed to provide a significant support for the research purpose during the pandemic.
    Click to visit github repo of the project
    Project Members: Sayeed Md. Shaiban, Abdullah Md. Sarwar and Arshi Siddiqui Promiti


Honor/Awards

Title: Champion of Banglalink-SDG Hackathon | Code for a Cause 2.0

Description: Noyona, an end to end product to help women with legal support by matching them with legal services, built using NLP based chatbot made with Google’s Dialogflow.

Click to visit github repo of the project
Personal Projects

Front-End Work

Algorithm 101

Here I try to list a number of popular algorithms starting with Sorting algos. The program uses a plugin architecture, so that anyone can write an algorithm and easily integret it to the system. A sorted file will be placed with the sorted numbers in the sorted folder.

lick to visit github repo of the project

© evrydaywannbe