Talking head research tops Altmetric chart

Share this on social media:

A research paper on synthesising realistic video-sequences of speech and expressions to mimic a particular individual has become the most widely-shared in the Altmetric 100's seven-year history.

The annual Altmetric Top 100 highlights research published in 2019 that has generated significant international online attention and discussion – from patents and public policy documents to mainstream media, blogs, Wikipedia, and social media platforms. Many of 2019's most polarising issues, from the increasing use of artificial intelligence to the climate emergency facing the population, are included in this year's chart. 

The top 10 articles of 2019 are:

  • Few-Shot Adversarial Learning of Realistic Neural Talking Head Models (Arxiv, May 2019)
  • Scientists rise up against statistical significance (Nature, March 2019)
  • Measles, Mumps, Rubella Vaccination and Autism (Annals of Internal Medicine, April 2019)
  • World Scientists’ Warning of a Climate Emergency (BioScience, November 2019)
  • A Style-Based Generator Architecture for Generative Adversarial Networks (Arxiv, December 2018)
  • Large-scale GWAS reveals insights into the genetic architecture of same-sex sexual behavior (Science, August 2019)
  • New elevation data triple estimates of global vulnerability to sea-level rise and coastal flooding (Nature Communications, October 2019)
  • Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial (British Medical Journal, December 2018)
  • The global tree restoration potential (Science, July 2019)
  • Civic honesty around the globe (Science, July 2019)

The Altmetric Top 100 typically features research from a variety of disciplines, written by authors from all corners of the globe. 

Other diverse subjects highlighted include the prorogation of Parliament, fake news on Twitter during the 2016 U.S. presidential election, mortality and survival in Game of Thrones, how Brexit will affect health services in the UK, and racial bias in an algorithm used to manage the health of populations. 

Egor Zakharov and Victor Lempitsky are two of the authors of the most discussed academic paper of this year, which looks at creating systems that can synthesise realistic video-sequences of speech and expressions that mimic a particular individual. Zakharov said: 'Our paper is inspired by the need to develop better telepresence solutions for augmented and virtual reality. The important component of telepresence applications is a realistic simulation of a person's appearance, which we try to perfect with the power of deep learning.'

Lempitsky added: 'Our technology can learn photorealistic head avatars from a handful of images of a person, using a new deep learning model trained on a large dataset of videos of celebrities.' 

The paper is the most widely shared in the Altmetric Top 100’s seven-year history.

Catherine Williams, COO at Altmetric, said: 'It’s fascinating to see the trends that shape the Top 100 list each year. In 2019, it’s clear that our current climate emergency and political polarisation are a matter of huge public concern and debate. This list demonstrates the critical role that research plays in those conversations.'

This year’s list features works published in 43 different journals, preprint servers, and government websites. The Harvard University authors appeared most often in the list (11 papers), while the journal Nature featured more than any other (12 times).

View the full list at https:/