Gain real-world industry experience through Break Through Tech AI at MIT | Massachusetts Institute of Technology News

This year, students from across the Boston metropolitan area are taking what they learned conceptually about artificial intelligence and machine learning (ML) and putting new skills into practice as part of experiential learning opportunities offered through Break Through Tech. We have had the opportunity to apply it to industrial projects around the world. AI at MIT.

Sponsored by the MIT Schwarzman College of Computing, Break Through Tech AI provides skills-based training, industry-relevant portfolios, and mentorship to undergraduate students to empower women and underrepresented genders in computing. is a pilot program aimed at filling the talent gap. Data to put metropolitan areas in a more competitive position for careers in science, machine learning, and artificial intelligence.

“Programs like Break Through Tech AI give us the opportunity to connect with other students and other institutions, reinforcing MIT’s values ​​of diversity, equity, and inclusion in the spaces we hold. It allows for learning and application,” says Alana Anderson. She is the Associate Dean for Diversity, Equity, and Inclusion in the MIT Schwarzman College of Computing.

The first group of 33 undergraduates from 18 schools in the Boston area, including Salem State University, Smith College, and Brandeis University, enrolled in an eight-week online skills-based course last summer for free 18-year courses. Monthly program has started. Fundamentals of AI and machine learning. The students then worked in small groups in the fall on his six machine learning challenge projects presented by MathWorks, MIT-IBM Watson AI Lab, and Replicate. Students spend over five hours each week meeting with their teams, teaching assistants, and project advisors, including monthly meetings at MIT, while juggling regular academic workloads with other day-to-day activities and responsibilities. Ta.

This assignment gave undergraduate students the opportunity to contribute to a real-world project being worked on by an industry organization and test their machine learning skills. Members from each organization also served as project advisors, providing encouragement and guidance to the team.

“Students gain industry experience by working closely with project advisors,” said Ord Oliva, director of strategic industry engagement in the MIT Schwarzman College of Computing and director of the MIT-IBM Watson AI Lab. “We’re piling up a lot of stuff,” he says. “These projects will be add-ons to your machine learning portfolio that you can share as examples of your work when you’re ready to apply for an AI job.”

Over the course of 15 weeks, the team combed through large, real-world datasets to train, test, and evaluate machine learning models in a variety of contexts.

In December, students celebrated their work at a showcase event at MIT, where six teams gave final presentations of their AI projects. The project not only allows students to gain experience in his AI and machine learning, but also “improve their knowledge base and skills in presenting their work to both technical and non-technical audiences.” It also helped, Oliva says.

For the transportation data analysis project, students will be trained in MATLAB, a programming and numerical computing platform developed by MathWorks, to create models that enable autonomous driving decision-making by predicting future vehicle trajectories. Did. “It’s important to realize that AI is not all that intelligent,” Brandeis University student Srishti Nautiyal said as she introduced her team’s project to the audience. Nautiyal, a physics and mathematics major with many companies making driverless cars a reality, said her team will also consider the ethical issues of technology in models to ensure the safety of passengers, drivers and pedestrians. He said that he is very enthusiastic about this.

Training models using census data can be difficult because it is often messy and full of holes. In the MIT-IBM Watson AI Lab’s Algorithmic Fairness project, the team’s most difficult task was cleaning up the mountain of unorganized data in a way that allowed them to derive insights from the data. This project aims to create a demonstration of fairness applied to real-world datasets, evaluate and compare the effectiveness of different fairness interventions and fairness metric learning techniques, and ultimately It may serve as an educational resource for data scientists interested in learning about and exploiting fairness in . It also aims to foster the practice of evaluating the ethical implications of machine learning models in industry.

Other challenging projects include an ML-assisted whiteboard for non-technical people to interact with off-the-shelf machine learning models, and a sign language recognition model to help people with disabilities communicate with others. It was. The team that worked on the visual language app set out to include his 50+ languages ​​in the model to increase access for millions of visually impaired people around the world. The team says similar apps currently on the market only offer up to 23 languages.

Throughout the semester, students demonstrated persistence and grit to get their projects across the finish line. The final presentation will conclude the fall semester, and students will return to MIT in the spring to continue their Break Through Tech AI journey and work on another AI project. Now, students can work with Google to tackle new machine learning challenges and further hone their AI skills with the goal of launching a successful career in AI.