Newcastle-based AI startup Literal Labs, which builds fast, efficient, and explainable logic-based AI models, has raised £4.6 million to grow its workforce and accelerate product development.
- Literal Labs, founded in 2023 as a spin-out from Newcastle University, develops AI models based on propositional logic rather than neural networks.
- The company’s approach uses formal logic rules to process data and make decisions, differing from neural networks that rely on layers of weighted connections inspired by the brain.
- This method reduces computational complexity, enabling faster processing and lower energy consumption. Because the models operate through explicit logical steps, their decisions are more transparent and easier to interpret.
- Recent MLPerf Anomaly detection benchmarks show that Literal Labs’ models can perform inference 54 times faster than traditional machine learning techniques while using 52 times less energy than comparable neural networks. The models also outperformed XGBoost, a widely used machine learning algorithm, with up to 250 times faster inference.
Details of the deal
- The investment was co-led by Northern Gritstone and Mercuri, with participation from Sure Valley Ventures, Cambridge Future Tech SPV, and several angel investors.
- The fresh funding will be used to grow the engineering team and develop the startup’s first commercial product, set to launch in 2025.
- This product will help customers with EdgeAI replace large, costly, and energy-heavy GPU algorithms, meet strict social or regulatory demands for AI explainability, and create more efficient AI models for battery-powered devices.
