States' Responding Behavior in Conflict: Asymmetric Response and Strategic Conflict Avoidance

State responses to external threats and aggression are studied with focus on two different rationales: (1) to make credible deterrent threats to avoid being exploited, and (2) to minimize the risk of escalation to unwanted war. Given external aggression, the target state's responding behavior has three possibilities: concession (under-response), reciprocation, and escalation. This study focuses on the first two possibilities and investigates how the strategic nature of crisis interaction can explain the intentional choice of concession or avoidance of retaliation. I build a two-level bargaining model that accounts for the domestic bargaining situation between the leader and the challenger for each state. The model's equilibrium shows that the responding behavior is determined not only by inter-state level variables (e.g. balance of power between two states, or cost of war that each state is supposed to pay), but also the domestic variables of both states. Next, the strategic interaction is rationally explained by the model: as the responding state believes that the initiating state has strong domestic challenges and, hence, the aggression is believed to be initiated for domestic political purposes (a rally-around-the-flag effect), the response tends to decrease. The concession is also predicted if the target state leader has strong bargaining power against her domestic challengers \emph{and} she believes that the initiating leader suffers from weak domestic standing. To test the model's prediction, I conduct a lab experiment and case studies. The experimental result shows that under an incentivized bargaining situation, individual actors are observed to react to hostile action as the model predicts: if the opponent is believed to suffer from internally driven difficulties, the subject will not punish hostile behavior of the other player as severely as she would without such a belief. The experiment also provides supporting evidence for the choice of concession: when the player finds herself in a favorable situation while the other has disadvantages, the player is more likely to make concessions in the controlled dictator game. Two cases are examined to discuss how the model can explain the choice of either reciprocation or concession. From personal interviews and fieldwork in South Korea, I find that South Korea's reciprocating behavior during the 2010 Yeonpyeong Island incident is explained by a combination of `low domestic power of initiating leader (Kim Jong-il)' and `low domestic power of responding leader (Lee Myung-bak).' On the other hand, the case of EC-121 is understood as a non-response or concession outcome. Declassified documents show that Nixon and his key advisors interpreted the attack as a result of North Korea's domestic political instabilities (low domestic power of initiating leader) and that Nixon did not have difficulties at domestic politics during the first few months of his presidency (high domestic power of responding leader).

Computational Analysis of Intelligent Agents: Social and Strategic Settings

The central motif of this work is prediction and optimization in presence of multiple interacting intelligent agents. We use the phrase `intelligent agents' to imply in some sense, a `bounded rationality', the exact meaning of which varies depending on the setting. Our agents may not be `rational' in the classical game theoretic sense, in that they don't always optimize a global objective. Rather, they rely on heuristics, as is natural for human agents or even software agents operating in the real-world. Within this broad framework we study the problem of influence maximization in social networks where behavior of agents is myopic, but complication stems from the structure of interaction networks. In this setting, we generalize two well-known models and give new algorithms and hardness results for our models. Then we move on to models where the agents reason strategically but are faced with considerable uncertainty. For such games, we give a new solution concept and analyze a real-world game using out techniques. Finally, the richest model we consider is that of Network Cournot Competition which deals with strategic resource allocation in hypergraphs, where agents reason strategically and their interaction is specified indirectly via player's utility functions. For this model, we give the first equilibrium computability results. In all of the above problems, we assume that payoffs for the agents are known. However, for real-world games, getting the payoffs can be quite challenging. To this end, we also study the inverse problem of inferring payoffs, given game history. We propose and evaluate a data analytic framework and we show that it is fast and performant.