PUBLICATIONS
Articles in Peer-reviewed Journals and International Conferences
Approximating Functions with Approximate Privacy for Applications in Signal Estimation and Learning
Naima Tasnim, Jafar Mohammadi, Anand D. Sarwate, Hafiz Imtiaz
Published in: Entropy
Publication date: May, 2023
Article number: 825
ALSNet: A Dilated 1-D CNN for Identifying ALS from Raw EMG Signal
K. M. Naimul Hassan, Md. Shamiul Alam Hridoy, Naima Tasnim, Atia Faria Chowdhury, Tanvir Alam Roni, Sheikh Tabrez, Arik Subhana, Celia Shahnaz
Published in: 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publication date: May, 2022
Pages: 1181-1185
M.SC. THESIS (ONGOING)
Supervisor: Dr. Hafiz Imtiaz, Associate Professor, Department of Electrical and Electronic Engineering (EEE),
Bangladesh University of Engineering and Technology (BUET)
Differentially Private Computation of Functions for Signal Processing and Machine Learning
Privacy has become an ever-growing concern in the modern age of information technology due to the massive increase in personal information stored in electronic databases. Simple anonymization of private information is not enough to ensure privacy, since the pattern of an individual's data is inseparable from their identity. An adversary observing the output of a machine learning algorithm can thus exploit the pattern of data to re-identify individuals. One state-of-the-art approach for solving the problem is differential privacy (DP). In essence, DP involves randomizing the outputs of an algorithm in such a way that the presence or absence of a single individual’s information within a database does not significantly affect the outcome of the algorithm. My current research involves developing a novel differentially private machine learning algorithm for regression problems that guarantees superior overall performance and privacy-utility trade-off over existing state-of-the-art approaches.
UNDERGRADUATE THESIS
Supervisor: Dr. S. M. Mahbubur Rahman, Professor, Department of EEE, BUET
SELECTED ACADEMIC PROJECTS
Attention Based Automatic Speech Recognition (ASR) with Sentiment Classification
Deep Learning Course Project (M.Sc.), 2022
Explored ASR with auxiliary classification objectives using attention-based networks on a small dataset with limited vocabulary
Performed effective speech transcription on the HarperValleyBank corpus, and showed the effect of model architecture on performance, utilizing BERT embeddings to improve performance of the downstream sentiment classification task
​
Relevant Skills: Python Report Colab Notebook
CNN Based Seizure Activity Classification Using Dominant IMFs of Band-Limited EEG Signals
Biomedical Signal Processing Course Project (M.Sc.), 2021
Used a well-established EEG dataset and exploited the knowledge of relevant EEG signal bands for seizure activity classification
Performed empirical mode decomposition (EMD) on five separate the band-limited signals and extracted dominant intrinsic mode functions (IMFs)
Used CNN based networks to classify non-seizure and seizure activities and compared performances on raw EEGs with band-limited EEGs
Relevant Skills: Python, MATLAB Report
Wearable and Low-Cost Device for Detecting Amyotrophic Lateral Sclerosis (ALS) Using EMG Signal
Biomedical Signals, Instrumentation and Measurement Course Project, 2020
Simulated circuit to obtain EMG signal from patients in Proteus
Used EMGLAB's dataset containing clinically obtained EMG signals of both healthy individuals and ALS patients
Developed "ALSNet": a novel dilated 1D convolutional neural network, to classify normal and ALS EMG, and compared performance with existing models
​
Relevant Skills: Python, Proteus Report Presentation
Copter Levitated Arm with Controller
Control Systems I Course Project, 2019
Implemented a copter arm with an appropriate controller consisting of three parts - an actuator, a plant and a sensor
Controlled the arm's levitation using feedback and feed-forward control mechanisms
​
Relevant Skills: Arduino, Python Report Reference (MITx 6.302.0x)
CNN Based Approach to Translate Bangladeshi Sign Language to American Sign Language (ASL)
Digital Signal Processing Course Project I, 2018
Collected image samples and developed dataset of Bangladeshi sign language (for digits 0-9)
Implemented CNN based model to translate Bangladeshi sign language to ASL
​
Relevant Skills: MATLAB
IC Based Morse Code Translator
Digital Electronics Course Project, 2019
Performed simulation as well as hardware implementation
Represented Morse code 'dot' and 'dash' with '10' and '01', respectively, at input end
Built a custom 8-to-256 decoder using multiple 4-to-16 decoder ICs and a custom 11-segment display using LEDs
Relevant Skills: Arduino, Proteus Report Project Folder
Other Relevant Projects and Assignments
(Undergraduate + Postgraduate)
Logistic Regression and Neural Network with One Hidden Layer in Python from Scratch (Machine Learning and Pattern Recognition Mid-term Assignment, M.Sc., 2022, Report)
K-means Clustering in Python from Scratch with and without PCA (Machine Learning and Pattern Recognition Final Assignment, M.Sc., 2022, Report)
IoT Based Class Attendance System Using RFID and Arduino (Measurement and Instrumentation Course Project, 2019, Report)
8-bit Simple-As-Possible (SAP) Microcomputer in Proteus (Microprocessor and Embedded Systems Mid-term Assignment, 2019, Project Folder)
Automatic Battery Charger (Power Electronics Course Project, 2019, Report)
​
Relevant skills: Python, Arduino, Proteus
