AI

Why are neural networks important?

Neural networks can help computers make intelligent decisions with limited human assistance. This is because they can learn and model the relationships between input and output data that are nonlinear and complex. For instance, they can do the following tasks.

RAGs

Retrieval-augmented generation (RAG) is a technique that combines large language models (LLMs) with information retrieval systems to improve the accuracy and relevance of AI outputs

Generative Adversarial Networks (GANs)

Two neural networks (generator and discriminator) work together to create realistic data, like images.

Two shot prompting

Two or more examples provided

Gen AI vs traditional AI

Unlike traditional AI that typically focuses on classification and prediction, GenAI focuses on generation — creating something novel based on a given input or prompt.

Self attention mechanism

a neural network mechanism that allows a model to prioritize the importance of different input elements when making decisions or predictions

Transformers

a neural network that learns context and meaning by tracking relationships in sequential data, like the words in this sentence.

Fine tuning

fine-tuning is the process of taking pre-trained models and further training them on smaller, specific datasets to refine their capabilities and improve performance in a particular task or domain.

Code analysis

using generative artificial intelligence (GenAI) technology to examine and evaluate code, identifying potential issues, suggesting improvements, and automating parts of the code review process by analyzing large amounts of code to find patterns and potential errors, thereby enhancing code quality and maintainability

Preventing Hallucinations

-Advanced prompting(instruct model to avoid false information) -Data Augmentation(RAGs) -Fine tuning

One shot prompting

A single example is given to clarify the task for the model

GitHub Copilot

-A code completion tool powered by OpenAI's Codex, designed to assist developers by suggesting code snippets, functions, and even entire algorithms based on the context of the code being written. -GitHub Copilot is trained on millions of publicly available code repositories and leverages deep learning models to predict and suggest code completions.

Transformers

-A specific neural network model models used for LLMs. -Transformers are able to learn context — especially important for human language, which is highly context-dependent. -Use a mathematical technique called self-attention to detect subtle ways that elements in a sequence relate to each other.

How LLMs work

-Architecture: LLMs are built on the transformer architecture, which uses self-attention mechanisms to understand the relationships between words in a sequence. -Training: LLMs are trained on large datasets from the web, books, academic papers, and other textual data sources. During training, they learn the statistical properties of language. -Fine-Tuning: After pre-training on general data, LLMs are fine-tuned on specialized datasets to improve performance on specific tasks.

Challenges and Ethical Considerations of GenAI

-Bias in AI: AI models can inherit biases from their training data, which could perpetuate stereotypes or discrimination. Deepfakes and Misinformation: AI-generated content can be used to create misleading information or harmful media. Content Moderation: Ensuring AI-generated content is safe and adheres to ethical standards.

Features of GitHub Copilot

-Code Autocompletion: Offers suggestions as developers write code, filling in functions, methods, or entire code blocks. -Function Suggestions: When you describe what you want to do in a comment, Copilot can suggest an entire function or method. -Documentation: Provides in-line comments and documentation to explain what a piece of code does. -Code Refactoring: Copilot suggests improvements to code quality, including better variable names or refactoring logic.

Large Language Models

Large Language Models (LLMs) LLMs are neural networks trained on vast amounts of text data to understand and generate human-like language. They can predict the next word or generate text given an initial input.

Machine Learning

Machine learning is a subset of Artificial Intelligence (AI) that enables computers to learn from data and make predictions without being explicitly programmed.

Neural Network

Neural networks a model based on the human brain's structure and function. A neural network consists of interconnected layers of nodes (analogous to neurons) that work together to process and analyze complex data. Neural networks are well suited to tasks that involve identifying complex patterns and relationships in large amounts of data.

Uses for Gen AI

-Content creation: AI-driven tools can automatically generate written content, such as articles, blogs, and reports. -Graphic design: Generative AI can do everything from creating original artwork to designing marketing materials like logos and banners. -Personalized marketing: Generative AI marketing tools can tailor marketing campaigns to each unique recipient by creating customized content that resonates with different segments of your audience. -Video game development: AI apps can generate realistic and interactive environments, character dialogues, and plot developments. That's how Starfield offered players nearly 1,700 planets to explore. -Production: These applications help with scripting, animating, and editing films by generating creative content and visual effects. -Music composition: Generative AI tools can compose music or assist musicians by generating chord progressions, rhythms, and entire compositions. -Code completion: AI tools for web development can speed up software development in multiple programming languages by suggesting context-appropriate code snippets, automating routine coding tasks, and reducing errors.

Code optimization use cases

-Generating optimized code for specific hardware or runtime environments -Suggesting code refactoring opportunities to improve performance -Automatically parallelizing code for better utilization of multi-core processors

Zero shot prompting

-No examples provided, must rely on pre-trained knowledge

Artificial Intelligence

-Refers to the simulation of human intelligence -Artificial intelligence (AI) is a set of technologies that enable computers to perform a variety of advanced functions, including the ability to see, understand and translate spoken and written language, analyze data, make recommendations, and more.

Types of Machine Learning

-Supervised Learning: Trains models on labeled data to predict or classify new, unseen data. -Unsupervised Learning: Finds patterns or groups in unlabeled data, like clustering or dimensionality reduction. -Reinforcement Learning: Learns through trial and error to maximize rewards, ideal for decision-making tasks

Prompt Engineering

-The process where you guide generative artificial intelligence (generative AI) solutions to generate desired outputs. -you choose the most appropriate formats, phrases, words, and symbols that guide the AI to interact with your users more meaningfully.

Security Considerations

-be aware of the potential for generating harmful or malicious content -misinformation/deepfakes -data breaches from datasets -data privacy from user inputs -malicious data injected into data sets

GenAI responsible use

-review and test AI-generated code before integrating it into a project. -be aware that AI tools might not understand the full context and can make mistakes. -be sure to respect licensing and intellectual property norms when contributing to open source

Variational Autoencoders (VAEs)

A generative model that creates new data by learning the probability distribution of the input data.

Prompt

A prompt is a natural language text that requests the generative AI to perform a specific task

Deep learning

A subset of machine learning that uses multi-layered neural networks to process and generate data.

AI Tooling

AI tooling refers to tools that use artificial intelligence (AI) to perform tasks such as writing, designing, and analyzing data. Some examples of AI tools include:

What are neural networks used for?

Computer vision, speech recognition, NLP, recommendation engines

Examples of Large Language Models

GPT-3, GPT-4(Generative Pre-Trained Transformer), Claude 3 family, BERT(bidirectional encoder representations from transformers)

Generative AI

Generative AI refers to a class of artificial intelligence systems that can create new content, such as text, images, audio, or even code, based on learned patterns from existing data.