Newsfeeds have proven to be powerful if controversial parts of social networks. Instagram has one. So do Twitter, TikTok, and Facebook. All have seen their share of negative press as of late, revolving largely around how fake news may affect elections, divide individuals on either side of a story, or spread misinformation about the coronavirus.
Obvious solutions to these issues might involve banning political ads from social media or showing stories from both (or more) opinions in everyone’s feed or even moderating certain types of content, such as news.
In this week’s Tech Tuesday, I will look at how one social network, Facebook, uses machine learning to power its news feed ranking algorithm with the hope that a view of its inner workings can help us to understand why it does what it does.
Behind the Curtain of Facebook’s News Feed
From a high-level, Facebook shows content to users that it believes the user will engage with. The goal is to maximize engagement, nothing more.
In a recent overview of the machine learning models that power their newsfeed, Facebook gives no mention of a couple of big items I would like to see companies like them tackle. These are:
- Diversifying content to show users both sides of a story; and
- Labeling news stories as unverified when appropriate.
Facebook’s personalization algorithm ranks content for over two billion people. They claim the objective is “to predict which content will matter most to each person to support a more engaging and positive experience”. This is among heaps of “overly promotional content or content from acquaintances who post frequently, which can bury the content from the people they’re closest to.”
Ranking is designed to solve this problem.
The algorithm looks for engagement signals among all the interaction data it collects from each of its users. From that data, the algorithm shows users more of what each user will likely engage with and rank this in ways that are predicted to be most engaging to them.
Facebook also runs user studies to ask users what content they feel is most engaging. In other words, Facebook is designing its algorithms to deliver irresistible content. And as far as hitting their stated objective they’ve done a great job.
But with their size and power, it’s worth asking ourselves if the big social networks are doing enough to address the societal issues they’re certainly helping to perpetuate, if not cause.
To learn more about this, I recommend diving into this article, written by Facebook’s engineering team. It’s surprisingly accessible and may help you form your own opinions about the ethical concerns surrounding the content shown to billions of people each day.
Building Inclusive Products Through A/B Testing
Here’s an article that addresses the novel approach to integrating product A/B testing and inequality measurement concepts from the field of economics by LinkedIn. The authors also discuss the methodology they have adopted for lowering barriers to economic opportunity in how different groups of members use LinkedIn products and, through a few choice case studies from the thousands of network A/B tests, provide examples of how it is helping reshape research and design practices at LinkedIn.
What Biden means for Big Tech—and Google in particular
Throughout his campaign, President-elect Joe Biden has been relatively quiet about the technology industry. In a revealing interview with the New York Times, Biden said that he wanted to revoke Section 230; suggested that he disagreed with how friendly the Obama administration became with Silicon Valley; and referred to tech executives as “little creeps”. But internet companies have also been among his campaign’s top 10 donors, technology industry insiders joined his campaign, and incoming vice president Kamala Harris has long-standing ties to Silicon Valley as a former district attorney in San Francisco. Here’s an article on what Biden means for big tech, and Google in particular.
Stabilizing Live Speech Translation in Google Translate
The transcription feature in the Google Translate app may be used to create a live, translated transcription for events like meetings and speeches, or simply for a story at the dinner table in a language you don’t understand. It’s been a wonderful tool but with early versions of this feature, the translated text suffered from multiple real-time revisions, which can be distracting. Recently, however, Google released an update to the transcribe feature in the Google Translate app that significantly reduces translation revisions and improves the user experience. Read more here.