BitsnBytes-Winter2024-2025

Bits & bytes Winter 2024-2025 5 What Is Generative AI? G enerative artificial intelligence (AI) is a subset of AI that refers to models and systems capable of generating new content, such as text, images, music, videos, and more. Compared to traditional AI, which primarily focuses on recognizing patterns and making decisions based on existing data, generative AI creates new data by learning from vast collections of data, and produces outputs that mimic human creativity. Generative AI uses large language models (LLMs) to perform tasks that involve natural language understanding and generation. LLMs are trained on large amounts of data to recognize patterns, structures, and relationships within the Continued on page 6 data. For example, text-based LLMs might be trained on books, articles, and websites. As LLMs process the data, they learn how words are typically used together, how sentences are formed, and the context in which certain phrases make sense, which is crucial for generating coherent text. Once the LLMs are trained, they can use what they learned to create new content, such as essays and stories. In short, LLMs are prediction engines that can guess the next words in sentences based on what came before them to finish sentences and paragraphs. LLMs are a key component of generative AI, enabling them to produce human-like text and perform a wide range of tasks. CONTENT CREATION ENHANCE DATA HEALTH CARE Applications and Examples of Generative AI • Text Generation: Certain models, such as ChatGPT by OpenAI, can write articles, generate poetry, create dialogue, and draft emails. • Image and Art Synthesis: Tools like DALL-E 3 and Midjourney can create new images from textual prompts, which can be useful for design and the entertainment industry. • Music: Some models, such as Suno, can compose music for creators. • Data Augmentation: Generative AI can create additional training data for machine learning models, improving their performance without requiring more real- world data. • Simulation and Modeling: It can generate realistic scenarios training autonomous systems, such as self-driving cars, in a virtual environment. • Drug Discovery: AI can generate molecular structures with potential therapeutic effects, accelerating the drug discovery process. • Medical Imaging: Generative models can enhance the quality of medical images and assist in diagnostics by creating synthetic images for rare conditions.