Last year at Bitjam we launched a small in-house machine learning project to coincide with the ACAVA Studios: Spode Works open studios event. ANNA – a learning algorithm that analyses poetry and audibly delivers it in an old-fashioned regional Potteries dialect – was the result!
Once we had our idea we sat down and had a discussion about the possible complexity of the project. A fully fledged bespoke neural network is quite a lot of work so we decided to try to find some existing neural networks to base our work off of. We found a Recursive Neural Network (RNN) designed to take text input and after a large number of training cycles we then tried to get ANNA to output some meaningful ‘learned’ poetry.
ANNA was a python script based on a simple RNN. We fed in around 200 pages worth of potteries dialect poetry aiming to produce some sort of meaningful poetry. Anna ran through about 500 recursive cycles of the input text per “epoch” of learning for a total of around 30 epochs. An epoch is essentially a single full training cycle.
The main challenge we faced was a shortage of data to feed into the RNN. So the next step for us was to source plenty of potteries dialect based poetry from poets past and present. First we tried to source as much poetry as possible from an online source, the main works we used were by Arnold Bennett. We then tried sourcing further poetry from Wilfred Bloor‘s sons Roger & Ian. Wilfred Bloor wrote over 400 Jabez tales in Potteries dialect (the Jabez character is a countryman living in the shadows of industrial Potteries). Finally, enter, Alan Barrett. Stoke-born writer, storyteller, poet, and actor who helped us train ANNA on the quirks and the peculiarities of the Potteries dialect. Thanks to the kind contributions of these people we managed to collect plenty of poetry that has been fed into ANNA.
The project was a lot of fun – especially since Alan Barrett remained on hand to help us deliver and chat about ANNA on the day of the open event! Our objective was only to take a playful look at neural networks and how they might be trained to learn local dialect. We didn’t exactly expect to achieve it, rather we were curious to see what the results might be. We got ANNA to recreate snippets of prose and dialect, and at times she successfully pieced together and understood some of the dialect.
ANNA was an example of Internet of Things – the interconnectivity of physical devices such as smartphones, WiFi modems and software – to the internet. We’ve got more projects coming up working with sensor data, redesigning systems, working with data from different types of sensors to create an interactive product, and we’re able to use our experience and knowledge of neural networks to complement these projects.
You can read the original ANNA blog post and interview with Bitjam Senior Developer Liam Mountford, in more detail here.
It’s fascinating to see the ways in which AI has exploded into the mainstream (hopefully we’ll avoid any Terminator-like “SkyNet” disasters!)
In particular, part one of our three part AI Series looks at using AI to crunch data with machine learning – generally described as a computer trained to learn and make recommendations based on data insights. These recommendations provide the ability to make accurate predictions and enhance system performances or personal experiences.
Machine learning has already been adopted in major technology-dependent sectors such as banking, healthcare, aviation and even space exploration:
“(ML) technology is expected to power future space exploration as it can handle huge data volumes, find patterns in planet image datasets, and predict spaceship condition.”
Satellites and space telescopes have already collected a large amount of data. Images provide the main source, but the challenge is how to identify information from the images. ML has become an effective technique for solving these problems.
Machine Learning and AI
Machine learning has progressed since its beginnings in pattern recognition. The theory that computers are able to learn without being programmed to perform tasks has evolved with the rise of AI, and since then research has been focused around proving that computers could actually independently learn from data.
Industries working with large amounts of data recognise the benefits of modern machine learning technology. By mining insights from this data companies are able to work more efficiently or gain an advantage over competitors.
Did you know that Machine learning can be applied to your existing systems in order to modernise and enhance your service and give you a competitive edge? It does not always require a brand new system that can be costly and disruptive. Another form of AI that can be adopted smoothly into existing systems is that of “Augmented Intelligence”.
AI: Augmented Intelligence
We are continually looking for projects to push the boundaries of using AI. When discussing AI we’re also referring to “Augmented Intelligence” as well as Artificial Intelligence. Augmented intelligence emulates and extends human cognitive function through the pairing of people and machines.
“Forward-looking companies and industry experts agree that augmented intelligence is the most effective way to maximize the value of AI”
Bitjam and AI
At Bitjam we recognise the need for software maintenance, updates and sometimes entirely new systems. We know that the idea of a complete overhaul of your IT-based systems is a concern, yet interoperability might cause long-term challenges in terms of business development. So how can Augmented Intelligence provide the solution?
Bitjam can help you connect with IoT devices to modernise existing services and delay the need for entirely new technology and data systems.
Last year we put our knowledge to the test and created ANNA – a machine learning algorithm that analyses poetry and audibly delivers it in an old-fashioned regional Potteries dialect. Part two of our three part AI Series this Friday will revisit ANNA and share with you the interesting results!
The final part of our three part AI Series next week will focus on the ways you can consider adopting AI for your own systems, how to approach decision-making and the ways in which Bitjam can assist you with the transition.
To learn more or to discuss any projects you have that Bitjam might have the solution to, you can email [email protected]
Recently at Spode it was the Acava Studios: Spode Works Open Studios event, and a chance for us to open our doors to the public and invite people to see what we’ve been up to lately here at bitjam.
We’ve introduced you already to ANNA, but if you didn’t see our last two blogs, ANNA is a learning algorithm that analyses poetry and audibly delivers it in an old-fashioned regional Potteries dialect. The open studios event was the chance for us to play around with ANNA, test her abilities and see if we could get a computer to understand and be able to reproduce Stokie dialect – something that’s famously difficult to do for non-natives!
Well, ANNA was, as NASA would say, a successful failure! She was designed from an idea that our senior developer Liam had had, as he has a background in machine learning from the work he did for his dissertation. So our objective was only to take a playful look at neural networks and how they might be trained to learn local dialect. We didn’t exactly expect to achieve it, rather we were curious to see what the results might be. We got ANNA to recreate snippets of prose and dialect, and at times she successfully pieced together and understood some of the dialect, but she sounded a bit more like actress Joanna Lumley than local storyteller and actor Alan Barrett (he helped us with the machine text input. You can read our Q&A with Alan about Stokie dialect here).
But what was amazing about this project was the ways in which is pulled people together from around the area, and got them talking about Potteries accent and dialect, from Keele University to artists and locals. It was a great celebration of the UK City of Culture bid and an opportunity to prove that Stoke does have digital creativity.
Special thanks to actor and storyteller Alan Barrett and local author Jason Snape for their contribution to ANNA, and to Roger & Ian Bloor for providing their Father Wilfred Bloor’s Jabez Tales (the Jabez character is a countryman living in the shadows of industrial Potteries).
In other news, our weather station “Thee Weather Duck”, will be going up in the Spode Works studios, giving artists and visitors the chance to tweet the weather in Stokie dialect. Something that came up in conversation time and time again during the ANNA project, is that young people are losing their accents, and are becoming very unfamiliar with old-fashioned Stokie dialect because of moving away to other areas for their studies, and the influence of the media and the rising popularity of standardised received pronunciation (RP). According to our recent Q&A with Alan Barrett, retaining dialect is a positive reflection on the community, and helps to establish and retain links to our cultural heritage and history. So maybe you’ve got children at university who need a reminder of home? Follow Thee Weather Duck on Twitter (@theeweatherduck) and RT the Stokie weather to them! The machine works with data collected from the weather station, and converts the results to Stokie dialect, for example “Iteside Temperature: Foetayn deegraze. If thees got chance, goo sunbeethin tidee!” (Translation: Fourteen degrees. If you’ve got chance, go sunbathing today!).
Our recent projects such as ANNA and Thee Weather Ducky are examples of “The Internet of Things”, a subject we’ve discussed in a previous blog post. The ‘Internet of Things’ is the interconnectivity of physical devices such as smartphones, WiFi modems and software, to the internet. We’ve got more projects coming up working with sensor data, redesigning systems, working with data from different types of sensors to create an interactive product, and we’re able to use our experience and knowledge of neural networks to complement these projects.