Among the benefits offered by algorithmic decision making and artificial intelligence – including revolutionizing speed, efficiency and ability to predict in a wide range of fields – Manish Raghavan is working on limiting related risk, while looking for the possibility of using technology to aid solve previously existing problems. Social fears.
“Ultimately, I want my research to head towards better solutions to long -lasting social problems,” says Raghavan, a professor of career development of Drew Houston, who is a member of the Mit Sloan School of Management department and the myth of Schwarzman College of Computing in the Department of Electrical and IT. Information and decision -making (LIDS).
A good example of Raghavan’s intentions can be found in his research on the apply of artificial intelligence in the employment process.
Raghavan says: “It is difficult to argue that the practices of employing in the past were particularly good or worth behavior, and tools learning on the basis of historical data inherit all prejudices and mistakes that people have made in the past.”
Here, however, Raghavan cites a potential opportunity.
“It has always been difficult to measure discrimination,” he says, adding: “Systems based on artificial intelligence are sometimes easier to observe and measure than people, and one of the goals of my work is to understand how we can use this better visibility to develop new ways of checking when the systems behave badly. “
Growing up in the San Francisco Gulf region, with his parents who both have scientific degrees in computer science, Raghavan says that he initially wanted to become a doctor. However, just before starting his studies, love for mathematics and computer science prompted him to follow the example of a family in the direction of computer science. After spending summer as part of bachelor studies at research at the University of Cornell under the guidance of Jon Kleinberg, a professor of computer science and computer science, he decided that he wanted to do a doctorate there by writing a master’s thesis on the “social influence of algorithmic decision making.”
Raghavan won awards for his work, including a reward in the scientific scholarship program for the National Science Foundation graduates, the Microsoft Research PhD Fellowship scholarship and the award for the doctoral dissertation of the Faculty of Computer Science of the University of Cornell.
In 2022 he joined the MIT Faculty.
Perhaps referring to his early interests of medicine, Raghavan conducted research on whether the determination of a very true algorithmic screening tool used in the segregation of patients with gastrointestinal bleeding, known as the Glasgow-Blatchford Score (GBS) scale, they are improved by the doctor’s advice .
“GBS study is more or less as good as in humans, but this does not mean that there are no individual patients or small groups of patients in whom GBS is wrong and doctors are probably right,” he says. “We hope that we will be able to identify these patients in advance, thanks to which the opinions of doctors will be particularly valuable to them.”
Raghavan also worked on the impact of internet platforms on users, taking into account the way social media algorithms observe the content chosen by the user, and then display more of the same type of content. Difficulty, says Raghavan, is that users can choose what they watch in the same way they could reach for a package of potato chips, which are of course delicious, but not very nutritious. Experience can be satisfactory at the moment, but it can make the user feel slightly ailing.
Raghavan and his colleagues have developed a user interaction model with contradictory desires – immediate satisfaction compared to the desire for long -term satisfaction – interacts with the platform. The model shows how you can change the platform design to ensure a healthier experience. The model won the Exemplary Applied Modeling Track Paper Award at the Association for Computing Machinery Conference on Economics and Computation conference in 2022.
“Ultimately, long -term satisfaction is important, even if you care only about the interests of the company,” says Raghavan. “If we can start gathering evidence that the interests of users and corporations are more convergent, I hope that we will be able to promote healthier platforms without having to solve conflicts of interests between users and platforms. Of course, this is idealism. However, I have the impression that a sufficient number of people in these companies believe that there is a place to make everyone happy, and they simply lack concept and technical tools to realize it. “
When it comes to the process of developing ideas for such tools and concepts of the best application of computing techniques, Raghavan claims that the best ideas come to his mind when he thinks about a problem for some time. He says that he advises his students to follow his example and put a very arduous problem for one day, and then return to him.
“The next day everything is often better,” he says.
When he does not solve the problem or teaches, Raghavana can often be found on the football field as coach Harvard Men’s Soccer Club, which he values very much.
“I can’t delay, knowing that I will have to spend the evening in the field, and this gives me a reason to wait for the end of the day,” he says. “I try to have things in my schedule that seem to me at least as important to work to put these challenges and failures in the right context.”
When Raghavan wonders how to apply computing technologies to best serve our world, he states that the most electrifying thing in his field is the idea that artificial intelligence will allow novel insight into “people and human society.”
“I hope,” he says, “that thanks to this we will be able to better understand ourselves.”