Applications of Artificial Intelligence in Science and Medicine | Hamzeh Ghorbani | TEDxUTMA

Quick Overview

Hamzeh Ghorbani details the application of Artificial Intelligence (AI) across scientific, medical, and industrial sectors, emphasizing its role in prediction and anomaly detection, which leads to optimizing processes, decreasing time, and reducing costs, as illustrated by examples like predicting supply chain distribution costs and detecting anomalies in well logging data.

Key Points: The idea of Artificial Intelligence was introduced by the British computer pioneer Alan Mathison in 1935, describing a computing machine that could move back and forth, read symbols, and write symbols. AI is defined as a computer system associated with human intelligence, involving learning, problem-solving, and decision-making, fundamentally differing from traditional programming where the program dictates the output based on data. An example application in management involves predicting Supply Chain Management Distribution Cost (SCMD), where AI helps optimize the best way to decrease distribution costs and find the best way quickly, decreasing analysis time. In industry (energy), AI is used for anomaly detection in well logging data, concluding that AI enables detection of abnormality, decreases time and cost for well logging, and reduces energy consumption. In medicine, AI, specifically using Convolutional Neural Networks (CNNs) in Computer Vision, predicts tumor delivery efficiency (DE) by nanoparticles, leading to the detection of breast cancer, and decreasing the time and cost associated with lab tests. The AI data pre-processing pipeline involves collection, cleaning, feature selection/reduction, handling outliers/imbalanced data, splitting data (train/test/validation), building the machine, evaluation, and final outcome prediction.

Context: This TEDxUTMA presentation by Hamzeh Ghorbani, a Researcher at ML, explores the practical applications of Artificial Intelligence (AI) across various fields: science (historical context), management (supply chain), industry (energy/well logging), and medicine (tumor delivery efficiency and cancer detection). The presentation contrasts AI methods with traditional programming and outlines the necessary steps for developing and applying AI models in these domains.