Amanda Christensen, ideaXme guest contributor, deepfake and fake news researcher and Marketing Manager at Cubaka, interviews Niall Donnellan, co-founder (with Cathal O’Gorman) of The History Machine Podcast, a project which uses a neural network to answer some of the “what ifs” of military history.
Amanda Christensen comments:
Fake news and media bias have been some of the defining elements of our most recent current events, from political elections to the current global crisis.
However, media bias and falsehoods are far from a new phenomenon; the recording of history has always favoured “the hand that holds the pen”.
Recent technological advancements, such as machine learning, neural networks and deepfakes have made it even harder to distinguish real from heavily biased, as well as “fake” news.
These same technological advancements, though, could potentially be the solution to detecting and removing bias from the recording of events as they happen and hence history.
In looking at this now, although neural networks can correct some bias, key caveats are of course that they are only as unbiased as a consistent pattern of information fed into them. “In our case, by keeping the information consistent, we can compare accurately. It is true that a neural network can correct for bias, but consistency is how this works”. Niall Donnellan, co-founder of The History Machine.
The History Machine
The History Machine is a neural network developed by “hobbyists” Niall Donnellan and Cathal O’Gorman designed to eliminate as much bias from military history as possible.
The History Machine analyses a database of wins, losses, casualties and other variables in the history of warfare, to present a more accurate depiction of history’s most famous commanders, generals, and leaders. Accuracy is of course also determined by whether the information recorded of past historians was accurate.
“We can only be as accurate as the source presented. For example, if the ancient historian Plutarch was not telling the truth, then we are analysing their lies in great detail”. Niall Donnellan.
It also allows them to test out hypothetical situations, pitting generals against one another to see which leader or army would reign supreme, as well as analysing fictitious armies and leaders, such as those in Game of Thrones.
Together, they present their results on The History Machine Podcast.
Niall Donnellan and Cathal O’Gorman
Niall Donnellan holds a Bachelors in Biomedical Engineering and a degree in Quality Engineering. He currently works in the biomedical sector as a Manufacturing Engineer at Cerenovus, working on The History Machine Podcast alongside Cathal in their spare time.
Cathal O’Gorman also holds a Bachelors in Biomedical Engineering as well as a Masters in Software Development, during which his thesis involved designing an AI which would determine what style and artistic movement a painting belonged to. Cathal is also the host of the Take your Data Points podcast, another podcast which uses a neural network to analyse and research data relating to the sport of hurling.
On this show we’ll hear from Niall about:
How he and Cathal came up with the idea for The History Machine Podcast. How the neural network works. The accuracy and degree to which his, and other neural networks, can remove bias. The potential use for neural networks in recording a less-biased version of major moments in modern history, such as the one we’re living in today.
The History Machine Podcast can be found on their website, as well as on Twitter and YouTube.
This interview is in British English
Credits: Amanda Christensen interview video, audio, and text.
Follow ideaXme on Twitter: @ideaxm
On Instagram: @ideaxme
On YouTube: ideaxme
Find ideaXme across the internet including on iTunes, SoundCloud, Radio Public, TuneIn Radio, I Heart Radio, Google Podcasts, Spotify and more.
ideaXme is a global podcast, creator series and mentor programme. Our mission: Move the human story forward!™ ideaXme Ltd.
Pingback: Optimising the Everyday with The Spatial Web - ideaXme