Artificial Intelligence (AI) enables computers to think, learn and solve problems like us humans. With unprecedented access to massive amounts of data and computing power, the pace of AI development has accelerated like never before.

The explosion of new data sources is revolutionizing the industries. The world is getting filled with information. Every day, millions of people use digital mass media generating over 2.5 quintillion bytes of data worldwide. Businesses are estimated to spend about $187 billion on data analytics in 2019. By 2021, insight-driven businesses are predicted to take $1.8 trillion annually from their less-informed peers. The amount of global data subject to data analysis is estimated to grow to 5.2 zettabytes by 2025. According to IBM, 90% of the data that exists in the world today was created very recently in the past few years.

“Big data” includes huge volume, high velocity and extensible variety of data. This data may be classified into three types – structured, semi-structured and unstructured. Fields of structured data are aligned side-by-side in fixed record lengths, with specific data fields appearing at static locations within each record. The data that organizations generate and use internally is structured. Examples of structured data are sales transactions, inventory and financial records, and flight and room reservation details. Organizations store structured data in relational databases such as Oracle and Microsoft SQL Server. Semi-structured data contains metadata or tags that are used to label the data without forcing it into a strict structure. The metadata provides additional information about the data entries. Semi-structured data is stored in text files using JavaScript Object Notation (JSON), Extensible Markup Language (XML), Comma Separated Values (CSV) and tab-delimited formats. Sources of semi-structured data include websites, Internet of Things (IoT) devices, apps and GPS trackers. Unstructured data does not have a recognizable structure, and so it is not typically stored in the traditional row-column databases. Unstructured data can come from a variety of sources, ranging from e-mail messages, documents, social media feeds, digital pictures, videos, audio transmissions, tweets, content from the web, sensors that gather information, and what not.

AI is a broad field; it includes Natural Language Processing (NLP) and Machine Learning (ML). NLP is the art and science that helps to extract information from text data. It enables computers to understand, interpret and manipulate human languages. NLP systems perform useful functions such as detecting languages, correcting grammar, converting speech to text, analyzing sentiment, categorizing documents and dynamically translating between languages. ML gives computers the ability to learn and improve from experience. It uses methods like statistics, neural networks and operations research to automate the building of analytical models. These models make it possible for computers to find hidden patterns and insights from large data sets. The main subfields of machine learning are supervised learning, semi-supervised learning, unsupervised learning, reinforcement learning and deep learning.

The dominance of data has caused a labor revolution in the finance industry. Quantitative analysts, or quants, have been assuming more significance while the traditional trading roles have been slowly disappearing. Quants analyze and interpret complex data using statistical and mathematical calculations. They use the results for forecasting and asset allocation. The amount of unstructured data in enterprises has been growing rapidly – often many times faster than the structured databases. In 1998, Merrill Lynch cited a rule of thumb that somewhere around 80% to 90% of all potentially usable business information may originate in unstructured form. The future appears to belong to unstructured data, and this data is too large to process with conventional tools. With the help of NLP and ML, quants can analyze large amounts of data to solve problems faster and more accurately.

Python is an open-source programming language that is easy-to-read and powerful. It is preferred for AI programming because the code is simple and clean. It is good for fast prototyping. It can be used to generate CSV files from raw data. These files can be imported into spreadsheets such as Excel. Python programs can also generate more complicated file outputs that can be consumed by machine learning algorithms. Along with NLP and ML, Python is becoming a favorite of many aspiring quants because of its utility in performing data analysis.

Another revolution is underway in the field of financial technology, or FinTech. Robo-advisors are technology platforms that offer investment advice suited to our financial goals. They manage the investments for us based on algorithms at a low cost. Conversational AI platforms, known as virtual assistants or chatbots, are transforming the retail banking experience. These assistants can help the customers to make payments, pay off loans and invest wisely. They provide the customers with real-time predictive analysis and update them on the state of their finances. They communicate through means the customers prefer, such as emails and instant messages.

Similarly, speech recognition driven by NLP and voice-to-data systems can transcribe and interpret quotes, conferences on quarterly results, payments and various other operations. Their outputs can automatically feed reports. The next generation of virtual assistants will respond to messages and make phone calls according to calendars. It is estimated that by 2020, about 85% of customer interactions will be managed by AI. In order to remain competitive, banks have to adapt their services by incorporating robotics.

This article on AI and finance will not be complete without mentioning Krishna C. Mukherjee. In the world of technology, there are few individuals who are as forward-thinking and innovative as Mukherjee. His illustrious professional career spans for more than three decades. He is a Microsoft veteran who has played a pivotal role in the development of Microsoft’s flagship products, including Microsoft Office and Windows. He has made highly significant contributions in the areas of software architecture, AI, distributed systems and cloud computing. He has created AI technologies that have streamlined and automated critical business processes. He has worked extensively in developing AI-powered systems that are widely used in an array of industries.

In the late 1990s, Krishna C. Mukherjee invented the “Intelligent Filing Manager” – a rule-based AI expert system that automates workflows and greatly improves the efficiencies of businesses. He created elegant software architectures that enable collaboration between cross-functional teams of an organization. Under Mukherjee’s leadership, Microsoft developed the Windows Presentation Foundation (WPF) and Extensible Application Markup Language (XAML) technologies that are very popular in the finance industry. The software developers work closely with quants and UX designers on FinTech projects. They work together collaboratively to create WPF and XAML applications that are information-rich and have minimum code complexity. These applications meet the security and high-performance demands of the finance industry.

In fact, Krishna C. Mukherjee instilled a new spirit in the finance industry. “Finance is a very important sector of our society,” he said. “I have devoted my time and efforts to create technology that impacts the lives of people who work in this industry.” He worked in leadership and directorial positions for reputed finance companies such as Citadel, Bloomberg and UBS. He provided strategic direction to these companies.

Nowadays, the investment decisions are no longer made using traditional passive strategies. Instead, they are made using a systematic and rule-based approach that Mukherjee pioneered. Numerous empirical studies show that we can achieve risk-adjusted returns higher than those obtainable from a conventional portfolio consisting of 60% stocks and 40% bonds. The rule-based strategies take advantage of anomalies related to fundamentals, market sentiments and price indicators. They offer a variety of solutions ranging from quoted funds (ETFs) that replicate indexes to more sophisticated offers with custom indexes to actively managed multifactor portfolios.

Moreover, Krishna C. Mukherjee has made significant contributions to quantitative finance. A most notable one is the architecture, design and development of the Bloomberg Valuation Service, or BVAL. In the past, prices of financial instruments were determined on the basis of market opinions and other comparable instruments. So, these prices were subjective and inaccurate. Financial decisions, based on these prices, were speculative in nature. This motivated Mukherjee to build a new pricing system from scratch. He designed BVAL to provide accurate valuations by using quantitative algorithms and AI. He created a scalable architecture for BVAL. He ensured that BVAL’s algorithms work off data that is of the highest quality. Today, BVAL prices millions of financial instruments across multiple asset classes. It provides the finance industry with rigorous, transparent and defensible valuations. By creating BVAL, Mukherjee enabled the finance industry to become objective and reliable.

Comments to: Applications of Artificial Intelligence in Finance

Your email address will not be published. Required fields are marked *