Algorithms or “algos” in deep learning are usually a finite set of instructions or steps followed to perform a calculation or task. Many useful algorithms are readily available today and data scientists spend much of their time choosing the right one or combination for specific machine learning applications.  The learning algo is what is used to train or find patterns in massive amounts of data related to a question or problem you are trying to solve. The output is a “model” that can be used to get predictions on data the model has not seen before.

In financial use cases, algo-trading or black-box trading uses a program that follows a defined set of instructions to place a trade. Some refer to this as high-frequency trading. The models used here are internally developed and are proprietary to the hedge fund or trading firm. Speed and execution of trades making only pennies of profit add up for large volume orders.