Blog article
See all stories ยป

Quo Vadis? Artificial Intelligence

In recent times the terms 'Artificial Intelligence' (AI), 'Machine Learning' (ML), 'Deep Learning' (DL) has gained much currency and established itself firmly in the context of 'Big Data' and 'Analytics'. Primarily these are software programs that analyse information and make decisions. Collectively let us call it the 'Brain'. That is how these are evolving to simulate how a human brain works. Can we leverage this fantastic potential for software application development that 'self-develop'?

The new frontier for AI, ML and DL is in software application development. More so in capital markets and risk management. In my view, a graduation from 'self-drive' cars. The software solutions today calculate net funding ratio or have the ability to process a swap deal over its life cycle. Given that the environment is ever changing, there is a constant need to upgrade the software to meet the dynamic business requirements. A case in point is the recent Basel III update on standardized approach. The banks using vendor solutions will now have to approach them for functionality enhancements. It is time to put a stop to making a beeline at the vendors doors each time something new comes up. 

The revised paradigm must be such that software applications should automate writing and execution of programs factoring in new evolving requirements. The requirements can be funneled into the 'Brain' to analyse, make decisions and provide processed extracts. This means the 'Brain' must 'remember' the current algorithm, analyse the new requirements, design and develop code for new fields, pointers to the right data, processing logic etc. This is the new universe. Oh! This will leave the army of coders with nothing to do in such a scenario, who have become Tofflerian illiterates? Well, the new breed will have to relearn to program not mundane tasks, but develop a creative 'Brain' that can project manage the SDLC lifecycle.

In the case of algorithmic trading that execute modelled trading strategies; 'AI' must do reverse engineering. Given the strategies, it must develop the models. 

7007

Comments: (0)

Vishwanath Thanalapatti

Vishwanath Thanalapatti

Analytics

Temenos

Member since

04 Jan 2018

Location

Canada

Blog posts

7

Comments

10

This post is from a series of posts in the group:

Capital Markets Technology

Front Office Trading Trends and Technologies...


See all