Promoting deeper learning and understanding in human networks

Peter Beshai

Laboratory for Social Machines, MIT Media Lab

The LSM team uses natural language processing, network science, machine learning, and user experience design to conduct analyses and build tools that promote deeper learning and understanding in human networks. 

Outcomes of prior work include:  

  • A study of the spread of false news that was the cover story of Science magazine and that Altmetrics ranked as the second most influential publication of 2018.
  • A tech-mediated coaching system, Learning Loops, for supporting kids’ narrative development that has been successfully piloted with hundreds of participants in collaboration with community organizations.
  • A partnership with the Commission for Presidential Debates to brief moderators of the 2016 debates, using large-scale analysis of social media to help to formulate questions for the US presidential candidates that reflected the national conversation.
  • A project, the Electome, to help US national newsrooms to analyze and report on the public conversation about 2016 election issues as revealed in social media. Post-election analysis of fragmented political networks… View full description

Laboratory for Social Machines, MIT Media Lab

The LSM team uses natural language processing, network science, machine learning, and user experience design to conduct analyses and build tools that promote deeper learning and understanding in human networks. 

Outcomes of prior work include:  

  • A study of the spread of false news that was the cover story of Science magazine and that Altmetrics ranked as the second most influential publication of 2018.
  • A tech-mediated coaching system, Learning Loops, for supporting kids’ narrative development that has been successfully piloted with hundreds of participants in collaboration with community organizations.
  • A partnership with the Commission for Presidential Debates to brief moderators of the 2016 debates, using large-scale analysis of social media to help to formulate questions for the US presidential candidates that reflected the national conversation.
  • A project, the Electome, to help US national newsrooms to analyze and report on the public conversation about 2016 election issues as revealed in social media. Post-election analysis of fragmented political networks and isolation of journalists led to one of Vice News’ most-viewed stories of the year. 
  • The creation of a media-technology non-profit, Cortico, that has brought a new kind of tech-scaffolded constructive conversation to thousands, now scaling across the country.
  • A coalition created in response to COVID-19, Beat the Virus, to deliver science- grounded public health guidance via social media influencers. LSM social media analytics guided the generation of over 600 million media impressions with no paid media. 
  • An established academic track record of over 160 peer-reviewed publications in human-machine communication and learning.

Team
We are an interdisciplinary team of researchers with backgrounds that include natural language and speech processing, machine learning + AI, interaction design, cognitive science, child development + learning, journalism and marketing.  We are committed to deploying our research through partnerships with external organizations such as the newsrooms, community organizations, schools, and libraries.