Iris is 2nd runner-up in AI Hackathon
The achievement underscores Iris’s capabilities in bringing real-world AI/ML solutions to its clients.
Iris Software recently displayed its prowess in artificial intelligence, the transformative technology that is becoming a game-changer across industries. An Iris team was ranked 2nd runner-up at the ‘30Hacks: Power of AI Hackathon’ in Noida, near New Delhi in May 2019. Two Iris teams were among the 27 that participated in the annual event, which is organized by the National Association of Software and Services Companies (Nasscom).
The innovative solution from Iris at the hackathon demonstrated how AI techniques could be used to validate conditions, assess outcomes and prescribe actions. Using a variety of ML models in a commonplace transaction such as an auto repair insurance claim, the solution was able to ascertain the type of vehicle and the extent of damage, prescribe the type of repair needed and assess its cost.
“The Nasscom win proves that the Iris AI/ML team can take on challenges in this area, and that it possesses the required technical skills to convert an idea into a workable solution,” says Abhishek Gupta, Head, Data Science (D&A Practice) at Iris Software, who led the teams that participated. When Iris teams score high at such competitions, it is testimony that they understand how to define a problem and how AI could be used to create the relevant solution.
Building on prior conquests
As it happens, Iris has consistently won AI awards. Its team was the second runner-up in the last edition of the Nasscom AI Hackathon held in November 2017, which was themed as ‘Machine Learning to Make India Smarter’. Gupta points out that while Iris’s Machine Learning capabilities are mainly in Natural Language Processing (NLP) and deep learning, the AI/ML team has been able to deliver to a diverse range of clients across domains and verticals.
Businesses in all organizations are joining data and analytics practitioners to leverage data science and artificial intelligence in their solutions. Iris has a mature team of data scientists who design innovative solutions with new tools and technologies across ML, NLP and advanced analytics.
Data science for report standardization
A large investment bank in North America confronted the need to process nearly 2,300 such reports in multiple formats, such as PDF, Excel or email. Going through communications in non-standard formats was making the task laborious.
The Iris Data Science team provided a solution to automate the reading of the fund reports, extract the required information with minimal manual effort and load it into a standard template by using ML and NLP technologies. By devising such a solution, Iris helped cut costs significantly for the client. The absence of manual intervention also removed the possibility of human errors in preparing data on fund performance. Further, it was possible to seamlessly move that data for other analysis.
Report rationalization
Organizations often undertake exercises to review existing reports and metrics with an objective to rationalize their reports and eliminate duplicates. In executing one such project for a large investment bank in North America, Iris used its Report Rationalization solution accelerator consisting of ML techniques and algorithms to create homogenous clusters of data from some 4,000 reports that were in multiple formats.
Next, a clean corpus of text was created by using text mining (TM) packages. The unsupervised machine learning algorithms were applied to the text corpus and duplicates were identified and clustered together. This resulted in rationalization recommendations of over 30% – it means the organization can now achieve its objective with 2,800 reports as against 4,000 reports.
It would have been impossible to achieve that if such a project were to be done manually at the rate of two reports per day by a risk analyst. By using data science, Iris could help its client avoid the huge costs, time requirements and errors that a manual effort would have meant.
Iris has shown considerable growth in AI/ML capabilities in the past year through the projects its Data Sciences team has undertaken. Gupta says that the aim is to create solutions that can help their customers realize the power of AI/ ML, optimize savings and reduce problem areas.
Since AI emerged in the 1950s, there have been more than 340,000 AI-related inventions, with a surge in innovation over the last five years, according to Nasscom. The AI market is projected to grow to $47 billion in 2020 and enterprises are rapidly adopting this technology wave to leverage most for their businesses, it states.
Contact
Our experts can help you find the right solutions to meet your needs.
Get in touch