Hi. I'm Goran Murić

I am a Postdoctoral Research Associate at the University of Southern California‘s Information Sciences Institute in the research group led by prof. Emilio Ferrara. I got a PhD in Engineering at TU Dresden at the Faculty of Electrical and Computer Engineering.

Most of my current research falls within the intersection of Machine Learning, Agent-based Simulations, Social Computing and Network Theory.

Some of the stuff I do.

Social networks

We build the agent simulation framework for studying dynamics in online environments. Details

Teamwork

We analyze team dynamics in online collaborative environments. Details

Spreading phenomena

I deal with the modeling of spreading failures within the networks and mitigating the risk. Details

Team optimization

Based on the friendship network, teams could be optimized to perform better. Details

Cascading failures

I model cascading failures within the network. I optimize the network so it becomes more robust. Details

Some videos I've made recently

SocialSim

I work on the COSINE project funded by DARPA’s SocialSim program. Together with the group of collaborators from the USC, Indiana University and University of Notre Dame, we build the first-of-its-kind cognitive agent simulation framework for studying multi-scale dynamics in online information environments. We build agents which can perform the set of actions analogous to the actions performed by humans in the real-world social networks. We model the agent behavior using data from GitHub, Twitter and Reddit. We apply the ML models together with the principles of human behavior and validate it through simulated experiments and empirical analysis.

Teamwork and productivity

How does the number of collaborators affect individual productivity? By analyzing the activity of over 2 million users on GitHub and Wikipedia, we discover that the interplay between group size and productivity exhibits complex, previously-unobserved dynamics: the productivity of smaller groups scales super-linearly with group size, but saturates at larger sizes. This effect is not an artifact of the heterogeneity of productivity: the relation between group size and productivity holds at the individual level. People tend to do more when collaborating with more people. We propose a generative model of individual productivity that captures the non-linearity in collaboration effort. The proposed model is able to explain and predict group work dynamics in GitHub and Wikipedia by capturing their maximally informative behavioral features.

Network resilience

I focus on network resilience against spreading failures. Spreading failures can propagate through the network and cause substantial damage or even the network breakdown. I model various spreading processes and try to identify the most influential spreaders within the network. Besides, I research cascading failures which are the phenomena prevailing in the networks loaded with traffic. Also, I try to implement the methods from the systems theory in the network analysis. You can find my paper on this topic here.

I use my findings to develop strategies to protect the network against the spreading failures. However, in some cases we want to do the opposite. For example, if we want to spread the news to the large audience of customers, we want to "infect" the social network on right places to do it most efficiently.

Teams optimization

The team will perform better if people in the team like each other. That is the main premise behind this project. We use the underlying network of social interactions within the large group to divide the group in optimal teams. Our algorithm solves this extremely complex optimization problem very efficiantly and timely. The researh part of the project is ongoing.

The solution could be used in education or in the companies where the team work is necessary.

Cascading failures

How much capacity is too much? Many networks are characterized by the capacity of its elements (nodes and links). A single failure could cause the traffic rerouting and further congestion and therefore more failures. Failures come in a form of a cascade.

Any system which can be modeled using an interdependence graph with limited capacity of either nodes or edges will be prone to cascading failure phenomena. I try to answer a big question: How to make networks more resilient against cascading failures?

You can ask me anything. Or just say Hi!


Find me on ...