You are here

Big Data and the Communications Gap

Write a message on a scrap of paper and try to throw it to someone to read across the room. When it comes to the act of communication, guiding that metaphorical paper and its message is one of life’s great challenges.

At The Ohio State University, electrical and computer engineering researcher Aleix Martinez recently won a three-year $900,000 grant from the prestigious Human Frontier Science Program (HFSP) to continue his work using Big Data to further bridge the communication gap between different people and cultures.

Through computers, Martinez is working to decipher the relatively unexplored realms of how social interaction and perception influence understanding between people.

The HFSP award funds his proposal, “Neural mechanisms underlying the visual analysis of intent.” Researchers at Caltech and the Hertie Institute for Clinical Brain Research in Germany co-authored.

The HFSP award is considered highly competitive, with less than 3 percent of all applications selected.

Martinez explains how his past research has explored the visual recognition of emotions, especially from faces, bodies and voice.

“Most of this work has focused on the perception of the emotional behaviors of individuals, and its underlying elementary perceptual processes, rather than on the specific system for recognition of intent,” he said.

His winning research proposal, however, seeks to better understand how visual perception identifies social interactions between people, as well as between animals and other organisms.

“Right now, we do not understand how the visual system can achieve this feat. We understand how facial expressions and some actions are recognized, but we know very little about how these and other percepts are integrated to determine which interaction is social and which type of social interaction that is,” Martinez said. “Our goal is simply to understand how visual perception solves the problem.”

The idea of perceiving social signals based on observing multiple cognitive functions, including the processing of visual features, action understanding, prediction and attribution of goals and mental states of others is already being explored via research, Martinez said. However, the fundamental computations and mechanisms involved in the perception and processing of social intent continue to remain unfamiliar.

“Equally unexplored is how the representations of social intent and semantics interact and influence each other,” he said. “Recent technical advances in machine learning, computer vision, and computational and cognitive neuroscience show that it is possible to extract a variety of variables that are important for inferring such semantic social information from visual stimuli, making studies of the perception of social intent possible. This includes specific social behaviors, the perception of the direction of intention of others, social roles taken in interaction, group context and the structure of social networks, and the explanation of proxemic cues, that is the relative spatial location of people in scenes.”

HSFP Research Grants enable scientists from different countries to collaborate on focused innovative projects that are expected to open new fields of investigation.

• Learn more about the HFSP at

• Learn more about Martinez and his work at the Computational Biology and Cognitive Science Lab: