Radcliffe Fellow’s Research Makes Hard Choices Easy

Research into computer organizational programs could make decision-making easier despite an increasing availability of big data, according to a presentation on Wednesday from Radcliffe fellow Shivani Agarwal.

Agarwal’s lecture, “Computers and Choice,” is part of the Radcliffe Institute for Advanced Study’s 2015-2016 Fellows’ Presentation series on research occurring at the Institute.

The lecture addressed problems consumers encounter when processing vast amounts of data because of continued growth in technology and consumer products.

“When we make choices today, one of the things we constantly face is a deluge of data, a deluge of choices, and a deluge of options, and somehow we are to make good choices in the face of all this complexity,” Agarwal said.

The goal of Agarwal’s current research is to find a way to categorize data to simplify the decision-making process.

“We want to find a way to organize items into categories based on their similarities in terms of how the users perceive them to make their choices,” she said.

Agarwal bases her research on machine learning, a type of programming that enables computers to engage in a combination of computational and statistical learning. She said machine learning is necessary to process immense amounts of data.

“Data is available at scales that no human mind can process, and therefore we need machines to organize and analyze this data,” Agarwal said.

Agarwal currently uses machine learning to analyze the choices people make to find “hidden” categories that people associate with a certain group of items. This categorization has the potential to help consumers more easily find desired items as well as identify which items they dislike.

This type of categorization has already proven successful, Agarwal said. Howard Moskowitz used this method to introduce a new type of pasta sauce in the early 1980s, “extra chunky,” that earned the Prego company $600 million in revenue.

Machine learning also has applications across many fields of study, especially the life sciences, according to Agarwal.

Agarwal said she hopes her research can have a similar impact.

“The idea is that there’s a lot of choice... we need some way to organize these choices,” she said.