Author Archive

Noshir Contractor and Kyosuke Tanaka presented their research at the 68th Annual Conference of International Communication Association in Prague, the Czech Republic: Tanaka, K., DeChurch, L., & Contractor, N. (2018, May). Origins of omission and commission errors in perceptions of group communication networks. Paper to be presented at the 68th Annual ICA conference, “Voices,” Prague, Czech Republic. Schultz, M., DeChurch, L., & Contractor, N. (2018, May). Communicating through space and over time. Paper to be presented at the 68th Annual ICA conference, “Voices,” Prague, Czech Republic.

“It’s Nicky Case!” is a cool website where you can play different kinds of games related to network concepts such as complex contagion and small-world networks. Moreover, there are other games for game theory and theory of voting. Check out the website (http://ncase.me/)!

Sid Jha will give a lightning talk “A Computational Platform to Evaluate the Ability to Perceive Social Connections” at the 2018 Computational Research Day on April 10, 2018. Moreover, Sid and Matt (both Undergraduate Research Assistants at SONIC) will present their posters then, respectively: Creating a Framework for Evaluating the Effectiveness of Various Search Strategies in the Small-World Phenomenon (by Matt) Network Acuity: Social Perceptions in a Small-World Experiment (by Sid) Both abstracts and posters are available: http://computational-research-day.s3-website.us-east-2.amazonaws.com/posters/

Hanna Wallach published a thought piece of what computational social science is, especially from her computer science point of view. Given computational social science in mind, She made points of differences between computer science and social science in terms of goals, models, data, and challenges:  Goals: Prediction vs. explanation — “[C]omputer scientists may be interested in finding the needle in the haystack—such as […] the right Web page to display from a search—but social scientists are more commonly interested in characterizing the haystack.” Models: “Models for prediction are often intended to replace human interpretation or reasoning, whereas models for explanation are intended to inform or guide human reasoning.” Data: “Computer scientists usually work with large-scale, digitized datasets, often collected and made available for no particular purpose other than “machine learning research.” In contrast, social scientists often use data collected or curated in order to answer specific questions.” Challenges: Datasets consisting of social phenomena raised ethical concerns regarding privacy, fairness, and accountability — “they may be new to most computer scientists, but they are not new to social scientists.”   She concludes her article saying that “we need to work with social scientists in order to understand the ethical implications and consequences of our modeling decisions.” The article is available here.    

A recent study by David Melamed and his colleagues examined whether the emergent structures that promote cooperation are driven by reputation or can emerge purely via dynamics. To answer the research question, they recruited 1,979 Amazon Mechanical Turkers and asked them to play an iterated prisoner’s dilemma game. Further, these participants were randomly assigned one of 16 experimental conditions. Results of the experiments show that dynamic networks yield high rates of cooperation even without reputational knowledge. Additionally, the study found that the targeted choice condition in static networks yields cooperation rates as high as those in dynamic networks. The original article is available here.

Recently Aaron Clauset and his colleague share their new study: “Scale-free networks are rare”. In this study, they found scale-free network structure is not so prevalent based on their statistical analyses of almost 1000 network datasets across different domains. In particular, their results indicate only 4% of the datasets showing the strongest-possible evidence of scale-free structure and 52% demonstrating the weakest-possible evidence. Additionally, this study has invoked intense conversations over Twitter. For instance, Laszlo Barabasi retweeted Aaron Caluset’s tweet, saying “Every 5 years someone is shocked to re-discover that a pure power law does not fit many networks. True: Real networks have predictable deviations. Hence forcing a pure power law on these is like…fitting a sphere to the cow. Sooner or later the hoof will stick out.” Link to the paper: https://arxiv.org/abs/1801.03400 Link to Barabasi’s retweet: https://twitter.com/barabasi/status/952920675592953856

Recently, Omodei, Brashears, and Arenas published a paper about describing a mechanistic model of human network recall and demonstrate its sufficiency for capturing human recall behavior based on experimental data. They found that human recall is based on accurate recall of a hub of high degree actors and also uses compression heuristics (i.e., schemata simplifying the encoding and recall of social information) for both structural and affective information. The original paper is here: https://www.nature.com/articles/s41598-017-17385-z

A recent study investigated a new way to identify the strength of ties. Using two different large datasets, the researchers found that for each pair of individuals, a bow tie structure of the network itself is strongly associated with the strength of ties between them that the researchers measure in other ways. The abstract of the paper is as follows: Understanding tie strength in social networks, and the factors that influence it, have received much attention in a myriad of disciplines for decades. Several models incorporating indicators of tie strength have been proposed and used to quantify relationships in social networks, and a standard set of structural network metrics have been applied to predominantly online social media sites to predict tie strength. Here, we introduce the concept of the “social bow tie” framework, a small subgraph of the network that consists of a collection of nodes and ties that surround a tie of interest, forming a topological structure that resembles a bow tie. We also define several intuitive and interpretable metrics that quantify properties of the bow tie. We use random forests and regression models to predict categorical and continuous measures of tie strength from different properties of the bow tie, including ...

Researchers at Annenberg School for Communication, University of Pennsylvania recently published a paper about “Network dynamics of social influence in the wisdom of crowds” in PNAS. They conducted an online network experiment where participants were asked to estimate numeric quantity (e.g., the caloric content) and tested how the accuracy of group estimates changes in different communication networks. They found that in decentralized networks, the group estimates were improved and in centralized networks, the accuracy of group estimates was undermined. Read the full article here.

“To figure out whether random AI can help people coordinate, Hirokazu Shirado, a sociologist and systems engineer, and Nicholas Christakis, a sociologist and physician, both at Yale University, asked volunteers to play a simple online game. Each person controlled one node among 20 in a network. The nodes were colored green, orange, or purple, and people could change their node color at any time. The goal was for no two adjacent nodes to share the same color, but players could see only their color and the colors of the nodes to which they were connected, so sometimes settling conflicts with neighbors raised unseen conflicts between those neighbors and their neighbors. If the network achieved the goal before the 5-minute time limit was up, all players in the network received extra payment. The researchers recruited 4000 players and placed them in 230 randomly generated networks. Some of the networks had 20 people controlling the nodes, but others had three of the most central or well-connected nodes already colored in such a way that they fit one of the solutions. (Each network had multiple solutions.) And some of the networks had 17 people and three bots, or simple AI programs, in charge of the nodes. In some networks, ...