- Oct 15, 2019
- Oct 10, 2019
Data, data and more data ...
Data is being collected, analyzed, and applied to solutions more widely and faster than ever before. Artificial Intelligence (AI) technology enables you to automate large amounts of data by applying them at registration speeds.
In the financial market, a share is the common denomination of a contract that, under certain conditions, gives the company the right or obligation to receive or provide assets or cash flows. A company in the financial market uses these actions to hedge risks when it operates in the markets. This calculation of risk and stock prices is a computationally intensive task and GPUs appear as one of the most important technologies to execute it.
Artificial intelligence (AI) is expanding rapidly and NVIDIA is positioning itself to gain more "market share" with a focus on new hardware and software to develop new products. The goal is to meet the demand of companies in the financial market that are increasing investments in artificial intelligence.
The company currently holds nearly 90% of the market for processors that are used for machine learning training tasks, also known as Deep Learning, and according to IDC, NVIDIA is among the leading cognitive software platforms focused on technological and business context changes.
The study also points out that one of the markets that will benefit most from this change is the financial market, with 28.6%, followed by the distribution and services sector with 20.8% and the public sector with 17.9% manufacturing with 14.1%.
Five years ago an operator needed to extract data and transfer it to specialized systems for advanced analysis and modeling. All this for the most important of your tasks: calculate the risks of the portfolio. In the past, risk calculations that required a lot of math were usually done in batches at night, making it difficult to react to real-time market changes.
With advances in video cards and deep learning, they can now perform data mining, development and modeling on a single, high-power computing platform with Kinetica and NVIDIA video cards.
Customers can perform complex on-demand queries without the need to move data between systems, and quantitative analysts can perform sophisticated scientific data work in the same database that stores all the information needed to guide negotiation decisions. This solves the challenge of data movement and makes it possible to use a simpler architecture for AI workloads.
With user-defined function capabilities accelerated by Kinetica video cards, customers can deploy a deep learning framework model such as TensorFlow, Torch, Caffe, or Spark ML using a simple API call. This enables quantitative analysts and other analysts to take advantage of the performance and benefits of video card parallelism without learning new programming languages.
On the other hand, performing the transactions involves figuring out how to get the best price for a security when you have a limited order certificate. If the future is a few hundred milliseconds ahead or there is only one minute as you negotiate ever greater amounts of a particular security, you want to know if you are getting the best price now in the near future.
By backing up your quantitative data in a deep learning environment, you can begin to understand where the millions of operations performed with this security go. You've trained with a tremendous amount of data and can then infer with this data in real time to see if you should trade now, in a few hundred milliseconds, in a second or a minute. This type of intelligence actually greatly increases the potential for algorithmic trading.
In the end, when the market closes, what you have is a great victory. Of the dice. And more data.