Mina Young Pedersen

• Galimullin, Rustam & Pedersen, Mina Young (2023). Visibility and exploitation in social networks. Mathematical Structures in Computer Science. ISSN 0960-1295. p. 1–30. doi: 10.1017/S0960129523000397. Full text in Research Archive Show summary

  Social media is not a neutral channel. How visible information posted online is depends on many factors such as the network structure, the emotional volatility of the content, and the design of the social media platform. In this paper, we use formal methods to study the visibility of agents and information in a social network, as well as how vulnerable the network is to exploitation. We introduce a modal logic to reason about a social network of agents that can follow each other, post, and share information. We show that by imposing some simple rules on the system, a potentially malicious agent can take advantage of the network construction to post an unpopular opinion that may reach many agents. The network is presented both in static and dynamic forms. We prove completeness, expressivity, and model checking problem complexity results for the corresponding logical systems.

• Pedersen, Mina Young; Slavkovik, Marija & Smets, Sonja (2023). Detecting bots with temporal logic. Synthese. ISSN 0039-7857. 202(3), p. 79–79. doi: 10.1007/s11229-023-04264-6. Full text in Research Archive Show summary

  Social bots are computer programs that act like human users on social media platforms. Social bot detection is a rapidly growing field dominated by machine learning approaches. In this paper, we propose a complementary method to machine learning by exploring bot detection as a model checking problem. We introduce Temporal Network Logic (TNL) which we use to specify social networks where agents can post and follow each other. Using this logic, we formalize different types of social bot behavior with formulas that are satisfied in a model of a network with bots. We also consider an extension of the logic where we explore the expressive power of including elements from hybrid logic in our framework. We give model checking algorithms for TNL and its hybrid extension, and show that the complexity of the former is in P and the latter in PSPACE.

• Baltag, Alexandru; Li, Dazhu & Pedersen, Mina Young (2022). A Modal Logic for Supervised Learning. Journal of Logic, Language and Information. ISSN 0925-8531. 31, p. 213–234. doi: 10.1007/s10849-022-09359-w. Full text in Research Archive Show summary

  Formal learning theory formalizes the process of inferring a general result from examples, as in the case of inferring grammars from sentences when learning a language. In this work, we develop a general framework—the supervised learning game—to investigate the interaction between Teacher and Learner. In particular, our proposal highlights several interesting features of the agents: on the one hand, Learner may make mistakes in the learning process, and she may also ignore the potential relation between different hypotheses; on the other hand, Teacher is able to correct Learner’s mistakes, eliminate potential mistakes and point out the facts ignored by Learner. To reason about strategies in this game, we develop a modal logic of supervised learning and study its properties. Broadly, this work takes a small step towards studying the interaction between graph games, logics and formal learning theory.

• Galimullin, Rustam; Pedersen, Mina Young & Slavkovik, Marija (2022). Logic of Visibility in Social Networks. In Ciabattoni, Agata; Pimentel, Elaine & Queiroz, Ruy J. G. B. de (Ed.), Logic, Language, Information, and Computation 28th International Workshop, WoLLIC 2022, Iași, Romania, September 20–23, 2022, Proceedings. Springer. ISSN 9783031152979. p. 190–206. doi: 10.1007/978-3-031-15298-6_12. Full text in Research Archive
• Pedersen, Mina Young; Smets, Sonja & Ågotnes, Thomas (2021). Modal Logics and Group Polarization. Journal of Logic and Computation. ISSN 0955-792X. 31(8), p. 2240–2269. doi: 10.1093/logcom/exab062. Full text in Research Archive Show summary

  This paper proposes different ways of modally defining properties related to the concept of balance in signed social networks where relations can be either positive or negative. The motivation is to be able to formally reason about the social phenomenon of group polarization based on balance theory. The starting point is a recently developed basic modal logic that axiomatizes the class of social networks that are balanced up to a certain degree. This property is not modally definable but can be captured using a deduction rule. In this work, we examine different possibilities for extending this basic language to define frame properties such as balance and related properties such as non-overlapping positive and negative relations and collective connectedness as axioms. Furthermore, we define the property of full balance rather than balanced-up-to-a-degree. We look into the complexity of the model checking problem and show a non-compactness result of the extended language. Along the way, we provide axioms for weak balance. We also look at a full hybrid extension and reason about network changes with dynamic modalities. Then, to explore the measures of how far a network is from polarization, we consider variations of measures in relation to balance.

• Pedersen, Mina Young; Slavkovik, Marija & Smets, Sonja (2021). Social Bot Detection as a Temporal Logic Model Checking Problem. In Ghosh, Sujata & Icard, Thomas (Ed.), Logic, Rationality, and Interaction 8th International Workshop, LORI 2021, Xi'an, China, October 16-18, 2021, Proceedings. Springer. ISSN 9783030887070. p. 158–173. doi: 10.1007/978-3-030-88708-7_13. Full text in Research Archive
• Pedersen, Mina Young; Smets, Sonja & Ågotnes, Thomas (2020). Further Steps Towards a Logic of Polarization in Social Networks. In Dastani, Mehdi; Dong, Huimin & Torre, Leon van der (Ed.), Logic and Argumentation Third International Conference, CLAR 2020, Hangzhou, China, April 6–9, 2020, Proceedings. Springer. ISSN 9783030446376. p. 324–345. doi: 10.1007/978-3-030-44638-3_20. Full text in Research Archive
• Baltag, Alexandru; Li, Dazhu & Pedersen, Mina Young (2019). On the Right Path: A Modal Logic for Supervised Learning, Logic, Rationality, and Interaction: 7th International Workshop, LORI 2019, Chongqing, China, October 18–21, 2019, Proceedings. Springer. ISSN 9783662602911. p. 1–14. doi: 10.1007/978-3-662-60292-8_1. Full text in Research Archive
• Pedersen, Mina Young; Smets, Sonja & Ågotnes, Thomas (2019). Analyzing Echo Chambers: A Logic of Strong and Weak Ties, Logic, Rationality, and Interaction: 7th International Workshop, LORI 2019, Chongqing, China, October 18–21, 2019, Proceedings. Springer. ISSN 9783662602911. p. 183–198. doi: 10.1007/978-3-662-60292-8_14. Full text in Research Archive

View all works in NVA

• Pavlova, Alexandra; Pedersen, Mina Young & Bernardi, Raffaella (2024). Selected Reflections in Language, Logic, and Information. Springer. ISBN 9783031506277. 221 p. Full text in Research Archive

View all works in NVA

• Pedersen, Mina Young (2023). Å oversette via esperanto. Mellom Tidsskrift for omsett litteratur. ISSN 2387-4880. 1, p. 99–106. Full text in Research Archive
• Pedersen, Mina Young (2021). Når maskiner lærer fordommer. SAMORA. ISSN 0801-6992. 3. Full text in Research Archive
• Pedersen, Mina Young (2024). Malicious Agents and the Power of the Few : On the Logic of Abnormality in Social Networks. Universitetet i Bergen. ISSN 9788230867006. Full text in Research Archive Show summary

  This thesis is situated at the intersection of formal logic, social science, information science and artificial intelligence research. It is devoted to the study of agents in multi-agent systems who act out of the ordinary or maliciously for their own gain. Analyzing these types of agents cannot only give us valuable information about their potential power to disrupt, but also about the safety of the systems in which they act. Specifically, we use social network logics to reason about malicious and abnormal behavior. Social network logics are logical systems in which a social network of connected agents is explicitly modeled. Although a social network logic is not necessarily based on modal logic, all logical systems presented in this thesis are. Four formal studies into malicious and abnormal behavior with social network logic are given. In the first, we build a temporal logic to model behavior associated with social bots, which are computer programs that act like human users on social media platforms. The logical models follow a social network as it evolves through a linear timeline. We propose to use model checking to detect whether, and if so, when there was social bot behavior in the network. A model checking algorithm is presented and it is shown that it runs in polynomial time. Then, we extend the logic with elements from hybrid logic. With this extended logic, model checking amounts to detecting not only whether and when, but also who in the network exhibited bot behavior. This model checking algorithm uses polynomial space. In the second study, we look at a social network where agents can post and share information—and show how an agent within the system can take advantage of the network construction to reach many agents with a controversial opinion. To examine such situations we present a logical system with dynamic and strategic extensions. Motivated by analyzing safety and possible exploitation in our system, we rely on a technique known as quantification over actions. We use this to formally study tactical actions from an agent's perspective. In the third and fourth studies, we extend logics from the literature to analyze the power of malicious and abnormal agents in existing systems. The third study builds on work on a particular network structure called structural balance. In this context we introduce new operators to view effects of friendships changing to enmities and vice versa. With these operators we can examine situations where single agents have the power to alter a global network structure. The fourth study explores a new type of agent in an existing model of threshold-limited influence. In this model, agents typically adopt a behavior or an opinion when a given number of their neighbors have adopted it, with respect to a given threshold. Inspired by existing work on different agent types, we discuss what it could mean to insert a malicious or abnormal agent into the diffusion. The result is an agent who when reaching the threshold can cancel out other agents' influence. With this thesis, we motivate logic as a unique and valuable formal tool to analyze social phenomena. We also show that formal characterizations of malicious agents can lay the groundwork for combatting them. Hopefully, this thesis will encourage further research on malicious and abnormal agents with social network logic.

View all works in NVA