Hello, this is
 Mohammad {Abuzar} Shaikh

PhD Candidate at University at Buffalo

About Abuzar

Abuzar is a doctoral candidate at State University of New York, Buffalo (UB). His research focuses on learning joint representations of data originating in multiple modalities to obtain a rich embedding useful for solving various downstream AI tasks. He is adept at applying Deep Learning (DL) in the domains of Computer Vision and Natural Language Processing and is open to discuss about opportunities in Deep Learning Research.


Jan'2018 - Current
Buffalo, NY, USA

Teaching Assistant

• Deep Learning CSE676 (Fall'2019, Spring '2020, Fall'2020)
• Advanced ML CSE674 (Spring'2018, Spring'2019)
• Intro to Machine Learning (ML) CSE574 (Spring'2018)

May'2019 - Aug'2019
Palo Alto, CA, USA

PhD Research AI, EY Labs

Deep Learning Research to overcome challenges in Document Intelligence using Self-Attention; Contextual-Attention; Generative Adversarial Networks

May'2018 - May'2019
Orlando, Florida, USA

AI R&D, RBC Innovation Lab

• Research using 3D-Convolution and conv-LSTM based computer vision and NLP models to solve challenges in Lip Reading. 
• Build ML models based on Time series analysis to forecast irregularly recurring expenses.

Jun'2017 – May'2018
Palo Alto, CA, USA

ML Research Intern, SAP Labs

Research using Recurrent Neural Networks (RNN) and collaborative filtering based Deep Neural Networks to solve challenges in recommendation systems.



Attention based Handwriting Verification
We propose a novel, attention-based model to compute probability of whether the given scanned handwritten evidences belong to same person


Self-Supervised Claim Identification
We propose a novel, attention based, self-supervised approach to identify “claim-worthy” sentences in a fake news article, an important first step in automated fact-checking.


Explanation based Handwriting Verification
We propose a method to generate explanations for the confidence provided by convolu-tional neural network (CNN) which maps the input image to 15 annotations (features) provided by experts


Hybrid feature learning - handwriting verification
We propose an effective Hybrid Deep Learning
(HDL) architecture for the task of determining the probability that a questioned handwritten word has been written by a known writer.


Writer Verification using CNN Feature Extraction
We propose an end-to-end learning method based on statistical features extracted on set-of-samples level as a step toward solving the writer verification problem which is about deciding whether two handwriting sources are identical given handwriting samples from the two sources.


Conf 2020

Document Intelligence

Coming soon

ICON 2020

Self Supervised Claim Identification

Implementation of the project is here

CSE676 Fall'2019


Handwriting Analysis

Implementation of the project is here

Where to find me!