Mohammed Aly's Projects
ArabiNizer is a state-of-the-art Arabic named entity recognizer (NER) leveraging the XLMR transformer model with an impressive testing accuracy of 95.00% and a remarkable testing F1-score of 88.00% on the PAN-X.AR subset from XTREME.
This project is an NLP (Natural Language Processing) application that classifies BBC news articles into different genres, including sports, politics, entertainment, business, and technology. The classification is done using two different techniques: LSTM and GRU.
An intelligent chess player based on searching algorithm the alpha-beta-pruning.
A Question Generation Application leveraging RAG and Weaviate vector store to be able to retrieve relative contexts and generate a more useful answer-aware questions
Welcome to Hakeem, your new Egyptian-Arabic Virtual Assistant! Hakeem is designed to simplify your life. With advanced features and natural language processing, it helps manage tasks, provides recommendations, and keeps you organized.
A simple house prices predictor trained on Boston housing prices dataset
Introducing JudgerAI - the revolutionary NLP application that predicts legal judgments with stunning accuracy! Say goodbye to the guesswork of legal decision-making and hello to unparalleled efficiency and precision.
Kevin is an advanced AI Software Engineer designed to streamline various tasks related to programming and project management. With sophisticated capabilities in code generation, Kevin can assist users in translating ideas into functional code efficiently.
A quick description of who am I and what I am doing
A detailed comparison between 3 different techniques (TF-IDF, Doc2Vec, and Sentence Transformers) for performing semantic search on a huge dataset
A comprehensive examination is conducted to assess the influence of homonyms in sentiment analysis, employing two distinct techniques: fixed embeddings (LSTM) and contextualized embeddings (DistilBERT).
A state-of-the-art Arabic part-of-speech tagger leveraging the XLMR transformer model With an impressive testing accuracy of 97.49% and a remarkable testing F1-score of 96.44% on the Arabic UD Treebank.
An Answer-Aware Question Generation Application Using Wikipedia as the Knowledge Source. Using T5-small and Instruction Fine-Tuning to generate a wonderful answer-aware questions.