Meta, a search engine powered by artificial intelligence (AI), is in the spotlight after news of its recent acquisition by the Chan Zuckerberg Initiative (CZI). The CZI has the implicit mission of “curing all disease,” unlike other charitable foundations that tend to be disease-specific.
By acquiring this powerful scientific search engine and providing it for free to all researchers, the CZI steps closer towards achieving its main goals: “To foster collaboration between teams of scientists and labs across multiple universities over long periods of time, to focus on developing tools that are geared toward eradicating diseases rather than simply treating them, and to improve and expand scientific funding writ large.”
In a blog post, Molyneux writes that the idea for his company stemmed from his attempt to address a common hurdle faced by many researchers: literature overload. “So I teamed up with my sister, Amy Molyneux — who happened to be an incredible developer with experience developing large-scale online platforms — and together we took on the challenge of building an online platform that would solve the problem of literature overload and allow people to stream and discover the literature in a more organic way,” wrote Molyneux. “Sciencescape, and eventually Meta, was born from that challenge.”
Meta’s impact on the research experience
In an interview with The Varsity, Michael Guerzhoy, a machine learning lecturer at U of T’s Department of Computer Science, noted Meta’s potential.
“There was enormous progress over the past ten or so years in terms of making the body of human knowledge searchable and accessible. You used to hear stories about graduate students working on a research idea for a long while before accidentally discovering that someone already published the same idea. Finding research papers, both old and new, that are relevant to your research used to be laborious.
“Meta aims to facilitate and, more excitingly, automate these processes,” said Guerzhoy.
For all students interested in pursuing a career in machine learning, Guerzhoy imparted the following advice:
“For people who are just starting university, my advice is to learn as much math as possible.” Guerzhoy notes that math is fundamentally what’s behind the “magic” of machine learning. To him, “it’s kind of amazing that you can use calculus and linear algebra to get insights from data.” He advises students to “take the most challenging math and statistics courses that you can handle, take an interest in them, and make sure that you deeply understand the material.”
Guerzhoy makes it clear that data is everywhere. “We are surrounded by data — election results, sports stats, financial news, etc.”
However, Guerzhoy makes it clear that theory needs to be practiced by students. “Once you have got a bit of a background in machine learning — through taking a U of T course, or perhaps through taking one of the several excellent online courses that are available because you couldn’t wait until your third or fourth year, start working on machine learning projects,” he suggests.
“In the past several years, a lot of excellent machine learning frameworks, such as Google’s TensorFlow, have become available,” notes Guerzhoy. “They make it relatively easy to build machine learning systems that perform very well. Right now, someone who successfully implemented an interesting machine learning system and can talk about it during a job interview would not have trouble getting an offer. Especially if you are considering graduate school, try to work on a course project or a summer project with a research faculty member.”
News about Meta has generated a lot of excitement and interest about AI among the U of T community.
“It’s great to see that U of T is still a hub for AI innovation,” says Matthew Scicluna, Statistics Lecturer at UTM and a machine learning enthusiast. “It’s very encouraging for future students who want to study machine learning, or use it in their business.”
A previous version of this article stated that “Meta has the implicit mission of ‘curing all disease.'” In fact, the Chan Zuckerberg Initiative (CZI) has this as their implicit mission.
