FACT NOT FAKE
SOLVING FAKE NEWS, FOR GOOD
Fake news has already undermined democracy, enabled large-scale fraud and incited violence. It also violates the fundamental human right to free choice, as it denies informed decision-making. Governments and the public are becoming increasingly concerned – and pressure is being applied to the biggest names in technology for a solution. Presently, the only solution is human review – slow, error prone and staggeringly expensive.
To solve fake news, Fabula AI developed (and patented) Geometric Deep Learning – the first AI technology able to learn from social networks. Our product has been trained to deliver unbiased authenticity scores for any piece of news, in any language, in milliseconds. Initial models are already achieving a best-in-class 93% accuracy (ROC AUC) and are still improving.
Fabula AI will deliver a better way to maintain trust in the world’s news: an independent, automated clearing house. To achieve this, we have assembled a world-class team of data scientists at the forefront of AI, together with successful entrepreneurs and seasoned technology leadership. With model accuracy proven, the focus of the Fabula team is dedicated to scaling up and productising its technology in order to open up commercial APIs for the world's social networks and publishers by the end of 2019.
The clearing house for the world’s news, in any language, on any platform – Fabula AI.
For more information, please contact us.
Automatically detecting fake news poses challenges that defy existing content-based analysis approaches. One of the main reasons is that often the interpretation of the news requires the knowledge of political or social context or "common sense", which current natural language processing algorithms are still missing. Recent studies have empirically shown that fake and real news spread differently on social media, forming propagation patterns that could be harnessed for the automatic fake news detection.
While Deep Learning approaches have achieved unprecedented success over a broad range of subjects (e.g. Computer Vision, Natural Language Processing, Speech Recognition, etc.), they are not suitable where the data is non-Euclidean (e.g. graphs and manifolds, such as social networks). Fabula has developed, and patented, Geometric Deep Learning which extends the proven Deep Learning techniques to graph/manifold structured data – ideally suited for working with social data.
Fabula’s product, and its unique approach to solving fake news, uses Geometric Deep Learning to establish the veracity of content by detecting the difference in propagation between true news and fake news. As this technology detects the spread pattern, it is language and locale independent; in fact, it can be used even when the content is encrypted. We also believe that such an approach, given it is based on the propagation pattern through huge social networks, is far more resilient to adversarial attacks.
Our models have already hit a best-in-class 93% ROC AUC and we observe reliable fake news detection after just a few hours of propagation – extremely encouraging initial results for Fabula’s fake news detection service.
For full details of our published results, see: https://arxiv.org/abs/1902.06673.
At Fabula AI, our mission is to make fake news a thing of the past, allowing platforms to quickly determine the authenticity of a news story through our APIs. We are a young and dynamic team of technology geeks backed by seasoned serial entrepreneurs and investors. We have developed (and patented) a novel class of deep learning algorithms which accurately predict fake news from their spread patterns over social networks.
As a Data Scientist or Data Scientist at Fabula AI, you will be responsible for developing novel, effective and scalable solutions to tackle very challenging problems. The ideal Data Scientist candidate will have expertise in statistical analysis, signal processing and machine learning. Excellent programming skills, creativity and independence are necessary for either position.
If you want to join us in ridding the world of the scourge of fake news, and you're a talented Data Scientist or Data Engineer, contact us.