SAN FRANCISCO, November 30, 2021 (Newswire.com) - Aible, the only enterprise AI solution that guarantees impact in one month, today announced the winners of the 2021 Real World AI Challenge held at the Berkeley AI Summit. The top scorers, which included a high school student and an MBA candidate, all used Aible and had no prior data science training. All of the contestants using AIble scored higher than the data scientists or participants using their own data science tools. Not one data scientist made it to the "top 10" in the contest.
In the UC Berkeley Real World AI Challenge, contestants competed to create a custom Artificial Intelligence (AI) models based on a real-world healthcare dataset to predict hospital patient readmissions. The challenge, sponsored by Aible and hosted as part of the Berkeley AI Summit, invited data scientists, PhDs, interested data novices, high school, and college students from all academic backgrounds to compete. The challenge: in two hours solve a real-world healthcare-related problem using Aible or the AI tool of their choice. Students analyzed a 56,000-patient dataset to determine patient readmission rates.
This year's event was judged by Michael Berman, CIO of California State University, and Gregory LaBlanc, Distinguished Teaching Fellow, University of California, Berkeley, Haas School of Business.
"It was pretty cool to see how quickly and accurately business students were able to come up with a machine learning model that was not only useful but that was designed to optimize real business value," said Berman.
Contestants Master AI in Minutes with AIble
The top score was recorded by Andrew Sova, an MBA Candidate at the Haas School of Business, who had no prior data science experience other than one hour of training on Aible. In a separate category for high school students, the winner was Neil Daniel, a 10th grader at the Thomas Jefferson High School For Science and Technology, recently rated the top high school in the country by U.S. News and World Report. The top-scoring data scientist, who has a PhD in physics, used his own data tools, but scored significantly below the contestants who used Aible. All category winners will receive a prize of $1,000.
"Competing against data science professionals was Intimidating and exhilarating, but Aible made the process of analyzing and predicting insights from the data set very clear and easy to do," said Sova. "Everyone will be able to use AI in the future, especially with tools like Aible."
Sova used Aible and finished the challenge in under 20 minutes after just one hour of training on Aible. That was well ahead of those with data science backgrounds who were competing using their own tools. The best data scientist submission did achieve higher scores in terms of data science metrics such as accuracy, but they delivered 20% lower results in the metric that mattered in the contest - expected economic impact of the right patient discharge decisions.
This continues to highlight Aible's core premise, highlighted by leading analysts and academic experts, that maximizing model accuracy does not lead to impact.
Data scientists were given four extra days to complete the challenge if they chose, but none decided to do so this year. Two years ago, Aible had sponsored the same contest at UC Berkeley, and in that case, only four out of 11 Data Scientists had managed to beat the score of the top high school student even after working on the problem for an extra five days.
Aible Open Challenge to Data Scientists
To enable more data scientists to check how well they did against the score of the Tenth Grade High School Student, Aible is issuing an open challenge where data scientists can access the training dataset and the use case write-up. Data Scientists would have to predict the outcomes for a test dataset, self-certify how long they worked on the problem, and specify the tools they used to achieve that result. The first data scientist to do better than the tenth grader on delivering economic impact as explained in the case write-up will get a $5,000 prize and bragging rights.
The results of the contest demonstrated that Aible was not only easy to use for a first-time user, but also produced very consistent results. The top scorers using Aible came up with the same answers, illustrating how Aible consistently produces highly reproducible results.
"The Aible user interface was straightforward; I felt like I could understand what I was doing and I still had the same power as a data science professional," said Neil Daniel, who took top honors in the high school student category of the Real World AI Challenge. "The sciences as well as the business sectors could really benefit from a simple and streamlined AI process like Aible."
Aible is #1 for Automated Machine Learning (AutoML) in Gartner's 2021 "Critical Capabilities for Cloud AI Developer Services' ' and is the only enterprise AI solution that guarantees impact in one month. The solution meets CIOs, CDOs, and business teams wherever they are on their AI journey — from AI-readiness, to data exploration with augmented analytics, to driving impact with optimized recommendations in end-user applications. With Aible, IT and data teams go from data to measurable revenue growth along with their key business stakeholders within 30 days. Discover real-world AI at www.aible.com.