BEN ELLIS

WHAT I DO

GENERATIVE ART
CREATIVE CODING
WEB DEVELOPMENT
DATA VISUALIZATION
INTERACTIVE DESIGN
PROTOTYPING

NOTES

Machine learning theory & Neural networks

3 MIN READ

Throughout the first 2 weeks of my Artifical Intelligence module, I have been learning about the theory behind machine learning and neural networks.

Defining machine learning - Tom Mitchell presents a key definition: "A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E."

FORMULA:
Data (input) —> guess = output (prediction) —-> feedback —> new guess (= repeat) etc.

We understood simple examples of Neural Networks, and how they are used to make predictions based on input data.

Neural Networks are made up of layers of nodes. Each node takes in multiple inputs, applies weights to them, sums them up, and passes the result through an activation function to produce an output.

Neural Networks Visualised

Routing & Building an API endpoint

4 MIN READ

Full-stack Module - Week 2

In our first 2 weeks of this module, we have been introduced to using node.js and express.js to build API endpoints and handle routing. To serve a web-page from our server we need to provide an index.html file in a public directory. We first call .use() on our app variable and pass express.static() to the use() function. Express.static() passes static files to our application. To pass a collection of files to the client side, we can crearte a 'public' folder which contains our static files such as index.html.

In order to communicate with our express server we need to create an API endpoint. This is done using app.get() method. The req parameter contains information about the request made to the server, such as query parameters, headers, and body data.

Routes - Routes are how data gets sent to and from the server and the client. A route can either be a GET or POST request. GET requests are used to retrieve data from the server, while POST requests are used to send data to the server.

Node Documentation
Express Documentation

Exploring a brief history of AI

3 MIN READ

■ (1936) Turing, "On Computable Numbers, with an Application to the Entscheidungsproblem" – computer decisions
■ (1956) Dartmouth Summer Project on AI, incl. Claude Shannon, John McCarthy, Marvin Minsky, funded by US military – trying to code concepts and generalizations
■ (1952) Christopher Strachey, Love Letters – plays with the idea of appearing human-like
■ (1957) Frank Rosenblatt, perceptron: electronic device able to "sense" and "learn from" environment – sensing and neural networks
■ (1964-67) Joseph Weizenbaum, ELIZA – attempt to refute the idea of appearing human-like
■ (1980) John Searle, Chinese room experiment – refuting the idea of appearing human-like
■ (1996) IBM's Deep Blue expert system defeats chess world champion Gary Kasparov – meanwhile, machine learning grows as a field
■ (2015) AlphaGo, machine learning-based, defeats Go master Fan Hui
■ (2017) "Attention is all you need": transformer architecture of machine learning
■ (2022) ChatGPT launched – machine learning

RECENT PROJECTS

HIGHLIGHTS

↳ LIGHTWAVES FESTIVAL 2025

CURRENT FOCUS

Building a web project for a local welding firm

1 MIN READ

Currently developing a full-featured website for a local welding firm. The project includes responsive design, interactive elements, and a clean, modern aesthetic to fit the firm's branding. The website currently utilises multiple html page with sections for services, products, contact information and images of the work they have done. The project is built using HTML, CSS and JavaScript, and is designed to be responsive and accessible. The project is still in development, but the current version and its updates can be viewed here:

The Local Welding Firm Website