Maryam KafiKang

• Developed, trained, and implemented a Neural Network controller, resulting in a substantial enhancement of RoboChair’s AI-driven Seat to Stand (STS) exercise equipment.

• Developed and executed a regression model with the primary objective of predicting children’s developmental milestones, utilizing extensive maternal drug usage history
• Skillfully reconfigured the original regression challenge into a classification problem, resulting in substantial improvements in predictive precision and accuracy.

• Vision Based Grasping”
• Designed an assistive neurorobotic device embedded with real-time Machine Learning pipelines to be used for disabled people
• Developed a Graphic User Interface (GUI) as the neural interface to collect brainwave data in synchrony with events in a hardware platform for reach-to-grasp tasks
• Preprocessing the collected brainwave data from the experiment and applying Machine Learning methods to classify signals into different grip types

• Detecting unforeseen Reovirus pattern in Turkey Genome sequences using Machine Learning
• Developing a Deep Learning model to detect and classify types of Reovirus

• Developed a deep learning model for DDI extraction, a challenging task in biomedical text mining.
• Applied Relation BioBERT, a powerful language model for biomedical text understanding.
• The developed model outperformed state-of-the-art models, demonstrating its effectiveness.
• This work is published in a top academic journal, further validating its significance.

• Performed data pre-processing including data cleaning, tokenization, and sentence embedding
using Fastai
• Developed a Deep Learning based classifier containing AWD-LSTM for text classification