What is AI?

Understanding what AI (Artificial Intelligence) is seems like a great place to start my journey on the path towards building my AI/ML (machine learning) knowledge.

artificial intelligence

/ˌɑːtɪfɪʃl ɪnˈtɛlɪdʒ(ə)ns/

noun

noun: artificial intelligence; noun: AI

1. the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Oxford Dictionary

AI is all about incorporating human intelligence inside machines.

Narrow Artificial Intelligence (or weak AI) are computer systems that can use human intelligence but have limitations in what they can do, for example Google Assistant and Siri. You can ask simple questions like “when is my next personal training session booked for” or “Call dad” or “what’s the weather forecast today?”.

Strong AI is the theory of a machine having an intelligence equal to humans. This is not something I am focused on.

That topics in AI that I intend to focus on are:

• Machine Learning (ML)
• Natural Language Processing
• Prompt Engineering
• Chatbot / AI Assistant development

Additionally

• Deep Learning
• Computer Vision – Images/AI creation
• Reinforcement Learning
• Ethical AI and Responsible AI

Machine Learning

Machine learning is a subset of AI and is the study of computer algorithms that improve automatically through experience. It includes two types of learning:

Unsupervised learning is the ability to find patterns or structure in data without our help. The more data we add the easier it becomes for the computer to see patterns that we may have missed. The computer may detect anomalies or cluster groups of similar data points together.

Unsupervised learning can be used in retail to segment customers into distinct groups based on similarities in their demographics, behaviours, preferences, or purchasing patterns. This can help a business in personalising their promotional material and marketing strategies. It can also improve the customer experience by delivering specific offers and messages to a particular segment.

Supervised learning is where we label data and give general guidance. The AI model algorithm needs the correct answer to succeed. The performance of supervised learning algorithms is typically evaluated using metrics like accuracy, precision, recall, F1-score, Mean Squared Error (MSE), etc.

Using historical sales data, promotional activities, seasonality and other relevant factors supervised learning algorithms can forecast future sales based on this information. Supervised learning can be used to detect fraudulent activities like identity theft and payment fraud. By teaching the computer what is and what isn’t an acceptable payment it can be taught to differentiate between what is and what is not acceptable.

Natural Language Processing (NLP – not to be confused with Neuro Linguistic Programming, which I do all the time!)

NLP involves the development of algorithms and models to understand, interpret, and generate language. It can support a business with managing their unstructured data. It can include videos and audio as well as text data. Commonly used areas of NLP include

Text classification – organise your business documents into categories e.g. HR, company documentation, budget etc.

Information Extraction – Find important information buried in large amounts of data without having to read it all yourself. Extract the information you need quickly.

Conversational agents – question answering
The ability to community with a user using natural language. It is very common to visit a website that has a customer service agent that can answer questions for you or a chatbot. Nowadays you can order food and request refunds using these agents.

Sentiment Analysis – You can analysis the sentiment in free-form text like reviews for example. You can search out unhappy customers, people who leave a negative or neutral review. Anyone who expresses anger. It can also work to help you pull out positive reviews and reviews from customers that are happy.

NLP fuelled by deep learning approaches have led to significant progress in tasks such as language translation, sentiment analysis, text summarization, and question answering.

Prompt Engineering

Prompt engineering is the area I am most interested in (at the time of writing this article). It focuses on our ability to craft effective prompts/instructions to guide an AI model like Chat GPT or Claude in providing the desired outcomes. The prompts created guide the behaviour of AI models by providing cues, constraints and context that will influence the information generated.

What I have been able to learn so far is that this isn’t a one time only process. Crafting prompts is an ongoing experiment, you have to be willing to create a prompt and then refine it based on the feedback provided. It becomes an iterative process as you explore strategies and best practices that will enhance the effectiveness of a prompt.

As I learn to design prompts I will be helping to shape the behaviour of an AI system to align with specific goals, preferences and ethical considerations. This is what will support me in developing chatbots/virtual assistants based on the needs of a business or collective.

Chatbot / AI Virtual Assistants

My aim in learning about things like NLP and prompt engineering is to help me develop chatbots and AI virtual assistants.

A chatbot (as mentioned above in conversational agents) is a program designed to simulate a conversation with a user. While they are popular in customer service there are also bots that provide guidance on a range of topics like fitness, recipes, books and more. You can ask a recipe chatbot to provide you with something to cook based on the ingredients in your cupboard, for example.

AI virtual assistants are more advanced in their abilities. They perform perform a wider range of tasks and can often handle complex interactions, manage calendars, provide personalised recommendations, and integrate with various services and applications.

Deep Learning

Deep learning is another subset of machine learning that deals with algorithms inspired by the structure and function of the brain’s neural networks. It is used in various domains such as computer vision, natural language processing, speech recognition, and more.

NOTE: The following explanations were created using Chat-GPT because I am not really focusing on them right now but may in the future:

1. Computer Vision:
   • Computer vision focuses on enabling computers to gain high-level understanding from digital images or videos. Research in computer vision encompasses areas such as object detection, image classification, semantic segmentation, instance segmentation, and image generation, with applications ranging from autonomous vehicles to medical imaging.
2. Reinforcement Learning:
   • Reinforcement learning involves training agents to make sequential decisions through trial and error interactions with an environment. Research in reinforcement learning explores topics like policy optimization, exploration-exploitation trade-offs, value function approximation, and transfer learning, with applications in robotics, autonomous systems, and game playing.
3. Ethical AI and Responsible AI:
   • As AI technologies are increasingly integrated into various aspects of society, there is growing recognition of the importance of addressing ethical concerns, fairness, transparency, and accountability in AI systems. Research in ethical AI and responsible AI focuses on developing principles, guidelines, and frameworks to ensure that AI technologies benefit society while minimizing potential harms and biases.

Now that my full-stack developer training is complete my aim is to use what I know about Flask, Django, API and databases to build chatbots and AI virtual assistants.

There is another way to go I discovered today. Rather than using all these tools I can use services like Microsoft Azure or the Google equivalent. Even IBM have a comprehensive learning package. I am still looking into exactly how I am going to put all this learning together.