Mohammed Alsobay

I’m exploring computational social science as a PhD candidate in the Information Technology group at MIT Sloan, where I am fortunate to be advised by Prof. Abdullah Almaatouq.

I worked in oil & gas, but felt that there were too many machines, and not enough people. I then worked in management consulting, and felt that there were a lot of people, but not enough technical work. Computational social science is my people-centric, technically-interesting middle ground.

When I’m not doing research, I may be found making obscenely large batches of banana pudding, trying to ski my way down New England’s icy slopes, or scuba diving while desperately hoping to see a turtle.

My two main research interests are:

👨 🤝 🤖 Human-AI interaction:

Social HAI: what do users learn from each other about algorithmic aids/agents, and how?

Beyond "speaking like a pirate": how do we steer LLM "personalities" in well-defined ways and measure the impacts on human-LLM interaction in competitive and cooperative settings?

AI-assisted social science: what role do algorithms play in theory formulation, and how can algorithmic tools (e.g. LLMs) enhance researcher productivity and research quality?

🧪 Digital experimentation:

LLMs as confederates: what new types of research questions are made possible by embedding LLM agents within human teams in behavioral studies?

More efficient and informative experiments: how do we integrate ideas from Bayesian optimization and "integrative" design to improve the quality of experimental social science?

Empirica: Contributing to the design of an open-source, low-fuss framework for running well-parameterized, interactive online behavioral experiments

updates

[2023-Oct-20]	Our paper on “The Effects of Group Composition and Dynamics on Collective Problem-Solving”, with Abdullah Almaatouq, Ming Yin, and Duncan J. Watts has been accepted and is forthcoming in Topics in Cognitive Science
[2023-Jul-17]	Presented at IC2S2 2023 in Copenhagen: [ Plenary talk] “Mapping the Effect of Altruistic Punishment on Cooperation” [ Poster] “The Effect of Social Information on Trust and Efficacy in AI-assisted Prediction” [ Tutorial] Creating multiplayer, interactive online experiments with Empirica, co-presented with James Houghton

[2023-Mar-23]	Presented a SAGE tutorial on Empirica (recording available)
[2022-Oct-8]	Presented a talk on “Integrative, high-throughput experimentation to explore social cooperation” at the “Measuring Belief Systems in Networked Communities” workshop at Princeton
[2022-Jun-7]	Passed my qualifying exams — thank you to Professors Abdullah Almaatouq, John Horton, Wanda Orlikowski, and Kartik Hosanagar for their support!
[2021-Nov-4]	Presented a parallel talk at CODE 2021 on “Collective Problem-Solving Across Tasks of Varying Complexity”
[2020-Sep-18]	Joined the Information Technology group at MIT Sloan as a PhD student!

publications

Task complexity moderates group synergy

Abdullah Almaatouq, Mohammed Alsobay, Ming Yin, and Duncan J Watts

Proc. Natl. Acad. Sci. U. S. A., Sep 2021

Abstract PDF Repo

Complexity-defined in terms of the number of components and the nature of the interdependencies between them-is clearly a relevant feature of all tasks that groups perform. Yet the role that task complexity plays in determining group performance remains poorly understood, in part because no clear language exists to express complexity in a way that allows for straightforward comparisons across tasks. Here we avoid this analytical difficulty by identifying a class of tasks for which complexity can be varied systematically while keeping all other elements of the task unchanged. We then test the effects of task complexity in a preregistered two-phase experiment in which 1,200 individuals were evaluated on a series of tasks of varying complexity (phase 1) and then randomly assigned to solve similar tasks either in interacting groups or as independent individuals (phase 2). We find that interacting groups are as fast as the fastest individual and more efficient than the most efficient individual for complex tasks but not for simpler ones. Leveraging our highly granular digital data, we define and precisely measure group process losses and synergistic gains and show that the balance between the two switches signs at intermediate values of task complexity. Finally, we find that interacting groups generate more solutions more rapidly and explore the solution space more broadly than independent problem solvers, finding higher-quality solutions than all but the highest-scoring individuals.
The Effects of Group Composition and Dynamics on Collective Performance

Abdullah Almaatouq, Mohammed Alsobay, Ming Yin, and Duncan J Watts

Top. Cogn. Sci., Nov 2023

Abstract PDF Repo

As organizations gravitate to group-based structures, the problem of improving performance through judicious selection of group members has preoccupied scientists and managers alike. However, which individual attributes best predict group performance remains poorly understood. Here, we describe a preregistered experiment in which we simultaneously manipulated four widely studied attributes of group compositions: skill level, skill diversity, social perceptiveness, and cognitive style diversity. We find that while the average skill level of group members, skill diversity, and social perceptiveness are significant predictors of group performance, skill level dominates all other factors combined. Additionally, we explore the relationship between patterns of collaborative behavior and performance outcomes and find that any potential gains in solution quality from additional communication between the group members are outweighed by the overhead time cost, leading to lower overall efficiency. However, groups exhibiting more “turn-taking” behavior are considerably faster and thus more efficient. Finally, contrary to our expectation, we find that group compositional factors (i.e., skill level and social perceptiveness) are not associated with the amount of communication between group members nor turn-taking dynamics.