48 Virtual Brochure – February 2024 4. Artificial Intelligence and the HE Sector–Mar 2023 Introduction Artificial intelligence (AI) is a branch of computer science that deals with the creation of intelligent agents, which are systems that can reason, learn, and act autonomously. AI research has been highly successful in developing effective techniques for solving a wide range of problems, including natural language processing, image recognition, and game playing. AI is rapidly transforming the world around us, and education providers are no exception. AI is being used to improve teaching and learning, research, and administration. It is also being used to create new products and services that can benefit society. ChatGPT is one AI tool that has dominated the news recently, though Google have just launched their competitor (Bard). These tools use GPT3 and LaMDA as their language models respectively. Both models are capable of generating text that is indistinguishable from human-written text, and both were used in the drafting of this paper. These two models (and ones similar) will drive AI developments for the next few years. We highlight the differences here to show that both models are very complex and well developed: 1. Purpose: GPT-3 is designed to generate text based on the input it receives, while LaMDA is specifically designed for dialogue applications, meaning it is optimised for generating responses to specific prompts in a conversational format. 2. Training data: GPT-3 was trained on a diverse range of text data, including web pages, books, and articles, while LaMDA was trained on conversation data specifically curated to mimic the types of conversations that people might have with Google Assistant. 3. Model size: GPT-3 is currently the largest publicly available language model, with over 175 billion parameters, while LaMDA is still in development and its size has not been publicly disclosed (though we expect it’s bigger). 4. Architecture: GPT-3 uses a transformer architecture with multiple layers of processing, while LaMDA uses a more complex architecture that involves multiple types of neural networks, including transformers and recurrent neural networks (RNNs). In layman’s terms, this means LaMDA is a more sophisticated tool (using more calculations in the process). 5. Applications: While both models can be used for a variety of applications, GPT-3 is more commonly used for tasks such as language translation, text completion, and text classification, while LaMDA is designed specifically for conversational AI applications, such as chatbots and voice assistants. Benefits of AI 1. Personalised learning: AI language models can be used to generate personalised learning materials for students, based
RkJQdWJsaXNoZXIy NTI5NzM=