Oxford Scientists Design A Tool To Train AI To Trade
A team of scientists at Oxford University has developed a first-of-its-kind GPU-accelerated limit order book simulator.
The JAX-LOB tool will teach AI to trade using Google's high-performance machine learning systems and financial data. The LOB simulator is one of a kind as it uses only GPU.
The use of GPUs allows artificial intelligence to skip several stages of data exchange, as it usually happens with CPUs. Thus, the speed of work is increased by seven times.
In traditional finance and the crypto space, LOB takes centrestage and is used constantly by professional investors. In order to train artificial intelligence to understand LOB, a large amount of simulation data is required. The level of development of the simulator determines the effectiveness of the model.
Oxford researchers assured that their product will help financial institutions to offer quality services. JAX-LOB can also manage market stability.
So far, JAX-LOB is still in the development stage and scientists are still exploring its possibilities. However, many are already predicting the product's success, convinced that JAX-LOB will be a revolutionary milestone in fintech.
