AI glossary · Fundamentals
Machine learning
In one sentence
Machine learning is the branch of AI in which a system learns to do a task from examples, rather than following hand-written rules, and gets better as it sees more data.
Also known as: ML · statistical learning · predictive modelling
What is machine learning?
In traditional programming, a person writes the rules and the computer applies them. Machine learning flips that round: you give the system lots of examples along with their outcomes, and it works out the rules itself. Nobody writes them; they come out of the data.
Put it to work
AI agents that look after your customers
From €49/mo + VAT
Voice, WhatsApp and email with the phone line included: agent, numbers and CRM in one fee.
See plans and pricingThat opens up problems where writing rules is hopeless. What exact rule separates a complaint email from a general enquiry? There are too many shades of grey. With enough labelled examples, a model learns to tell them apart with a high hit rate, though never a perfect one.
It underpins almost all of today’s artificial intelligence, including the deep learning that language models run on.
Related to this term
UTIO agents
- Available 24/7 in 26 languages
- Set it up yourself for €0: built from your website
- 14-day free trial, no card needed
From €49/mo + VAT
See plans and pricingHow it works
- Data. You prepare a dataset: representative examples of the problem, cleaned up and, for supervised learning, labelled with the right answer.
- Training. An algorithm adjusts a model so its predictions match the right answers as closely as possible. See model training.
- Evaluation. You test the model on examples it never saw in training. If it only does well on the ones it already knows, it has memorised rather than learned. That’s called overfitting.
- Deployment and monitoring. The model goes to work on real data. Because the world changes, you keep an eye on whether its accuracy drifts down over time.
Example
A garage in Bristol gets hundreds of messages a week through its website and WhatsApp. It wants to see at a glance which are MOT bookings, which are quote requests and which are complaints.
With a classification model, every message arrives already tagged. The model learned from thousands of previously sorted messages and now tags new ones. When it gets one wrong, someone corrects the tag, and those corrections feed into the next version of the model.
What it means for your business
You don’t need a data science team to benefit from machine learning. You already use it whenever a system transcribes a call, understands a message or works out what an email is about.
UTIO’s agents run on models that are already trained; you don’t train anything. What you bring is your business knowledge, as in the support vertical, where the agent answers from your documentation and opens a ticket when it can’t find the answer. If you’re a developer, the AI API includes text for replies, summaries and classification, with the model tier of your choice.
Frequently asked questions
What are the types of machine learning?
Three main ones. Supervised learning uses examples labelled with the right answer (emails marked spam or not). Unsupervised learning looks for groups or patterns in unlabelled data (customer segmentation). Reinforcement learning learns by trial and error with rewards and penalties, as in games or robotics.
What's the difference between machine learning and deep learning?
Deep learning is a kind of machine learning that uses neural networks with many layers. It shines on complex data like text, speech and images, but needs more data and more computing power than other methods.
Do I need lots of data to use machine learning in my business?
To train your own model, usually yes, and well organised. But most small businesses train nothing: they use models a provider has already trained and apply them to their own data, which takes far less information.
Related terms
- Artificial intelligence
Artificial intelligence (AI) is the branch of computing that builds systems able to do tasks we associate with human intelligence, such as understanding language, recognising images or making decisions, by learning patterns from data.
- Deep learning
Deep learning is a type of machine learning that uses neural networks with many layers. Each layer picks out more abstract features of the data, which lets it handle text, speech and images without hand-written rules.
- Dataset
A dataset is an organised collection of examples, such as text, audio, images or records, used to train, fine-tune or evaluate an AI model. Its quality directly shapes what the model learns.
- Model training
Model training is the process in which an AI model learns from a dataset: it's shown examples, its errors are measured and its parameters are adjusted, over and over, until its outputs are good.
- Algorithm
An algorithm is a finite sequence of well-defined steps for solving a problem or completing a task. In AI, the algorithm is the method that learns from data; the model is what that method produces.
Updated on