A new machine-learning club at Los Altos High School has propelled local students into the still budding world of self-driving cars.
Since forming the Electric Dreams club in March, freshmen Ravi Krishna and Maxwell Liu have led more than 30 peers in building an autonomous model car.
If successful, the converted radio-controlled car, dubbed “ShadowCart,” will navigate autonomously through the use of an advanced neural network, a form of artificial intelligence. Already ShadowCart can recognize a banana, a fork, a stuffed dog and a toy bird. The car moves in specific directions depending on each object’s identity. When it detects the bird, for example, it automatically turns left. When it sees a banana, it pulls forward.
Once complete, ShadowCart should be able to follow a person at a roughly constant distance and navigate around obstacles as necessary. Electric Dreams members expect to finish the car by the end of the school year and then spend the following two years assembling a self-driving golf cart, “NeuroCart.”
These projects, Krishna and Liu believe, will give members of Electric Dreams an advantage once they enter the job market. Machine learning extends far beyond self-driving vehicles, to speech recognition software, virtual personal assistants, medical diagnostic tools, advertising and finance. The world’s largest tech companies have invested heavily in its use.
Krishna and Liu are not aware of any other projects that involve high school students in building self-driving vehicles or other applications of machine learning.
Krishna acknowledged the advantages of starting a machine-learning club in Silicon Valley. Google’s Waymo, he said, has helped generate excitement with the self-driving fleet it began to test on the streets of Mountain View in 2009. In addition, students who grow up in Silicon Valley are often exposed to computer science at a young age.
Still, Krishna and Liu estimate that roughly one-third of Electric Dreams’ members had no experience in computer programming when they first joined the club.
“That’s who we think this club is really for,” Krishna said. “It’s for people who are interested in this field, who maybe don’t know a lot already.”
Spreading the word
Electric Dreams hopes to reach students beyond Los Altos as well. Krishna noted the possibility of expanding into neighboring Mountain View High School. Jeff Harding, superintendent of the Mountain View Los Altos Union High School District, expressed enthusiasm about the club’s capacity to engage students.
“Their excitement for learning is contagious, as more students get involved with the technology,” Harding observed.
Electric Dreams sees the opportunity to replicate its work outside of Silicon Valley as well. Liu has proposed creating kits with the base parts to build a prototype similar to ShadowCart. Students at any school could use the kit and possibly start their own machine-learning clubs.
“The goal is that students can empower themselves so that, if they hear about this or if they’re interested in it, then they can just pick up this kit and start working on it,” Krishna said.
Electric Dreams has pursued collaboration with high-tech companies as well. The club worked with staff from Santa Clara-based NVIDIA to develop a formal proposal to build ShadowCart. NVIDIA produces the artificial intelligence supercomputer that controls ShadowCart.
Electric Dreams members believe that the club could attract support from other Silicon Valley companies concerned about educating the work force of the future. National data show that by 2020 there will be demand for 1.4 million computer science-related jobs and only 400,000 computer science graduates to fill them.
In the self-driving auto industry, the going rate for talent has surged as companies compete for labor. In an interview with the tech news site Recode last year, Sebastian Thrun, who formerly ran Google’s autonomous car program, put the rate at a staggering $10 million per employee.
“There are just so few people in the field right now,” said Krishna, adding that “you can see something in (machine learning and deep learning) technology – that it’s really going to change a lot of things.”
For more information, visit electricdreamslahs.wordpress.com.