Introduction
Assumption-Based Argumentation (ABA) is a prominent structured argumentation formalism used to model reasoning through assumptions, rules, and conflicts. Its extension, ABA+, incorporates preferences to handle scenarios where certain assumptions are prioritized over others. However, ABA+ has historically faced two significant challenges: the lack of a faithful abstract representation (as standard Dung-style graphs or SETAFs are insufficient) and the "empty outcome" problem, where valid reasoning results (extensions) are not guaranteed to exist.
This paper introduces Hyper Argumentation Frameworks (HYPAFs), a novel abstract model that employs set-to-set attacks to fully capture the dynamics of ABA+. Furthermore, the authors propose the $Θ$-operator, a cautious defense mechanism that ensures the existence of grounded and complete extensions, resolving the long-standing issue of non-existent outcomes in preference-based argumentation.
Hyper Argumentation Frameworks (HYPAFs)
To represent the complex interactions in ABA+, the authors move beyond simple point-to-point or set-to-point attacks. In a HYPAF, the attack relation is defined between sets of arguments ($S \to T$). This level of abstraction is necessary because preferences in ABA+ can lead to situations where an attack is reversed, effectively making one argument conflict with an entire set of assumptions.
The paper proves that any ABA+ framework can be translated into a HYPAF while preserving its underlying semantics. This provides the first truly abstract, graph-based visualization for ABA+, allowing researchers to analyze preference-based conflicts through a structural lens.
The $Θ$-Operator and Cautious Defense
A core discovery of this work is the failure of the "Fundamental Lemma" in the context of HYPAFs. In traditional argumentation, if a set of arguments defends an argument $a$, $a$ can be safely added to the set. In HYPAFs, however, adding a "defended" argument can inadvertently introduce new internal conflicts within the set.
To address this, the authors introduce the $Θ$-closure operator. Unlike traditional defense, $Θ$ is "cautious": it only includes an argument in a set's closure if its inclusion is justified and does not violate the set's consistency. This operator forms the basis for new $Θ$-complete and $Θ$-grounded semantics.
Key Results and Complexity
The introduction of $Θ$-semantics yields several critical theoretical breakthroughs:
- Guaranteed Existence: $Θ$-complete and $Θ$-grounded extensions are guaranteed to exist for every HYPAF and, by extension, every ABA+ framework. This ensures that the reasoning system always produces a valid outcome.
- Uniqueness: The $Θ$-grounded extension is unique, providing a single, skeptical set of "winning" arguments.
- Computational Efficiency: Complexity analysis reveals that finding the $Θ$-grounded extension is often more efficient than traditional methods. Because it can be computed via simple iteration of the $Θ$ operator, it sits lower in the polynomial hierarchy than many comparable problems in structured argumentation.
Conclusion
By redefining the abstract foundations of ABA+ through HYPAFs and the $Θ$-operator, this research provides a robust solution to the "empty outcome" problem. While the translation from ABA+ to HYPAF is not always bijective—meaning some specific preference nuances might be abstracted away—the framework offers a powerful tool for reliable automated reasoning. Future work will focus on applying these "set-to-set" concepts to other formalisms like ASPIC+ and developing efficient algorithmic solvers for large-scale knowledge bases.
Summary updated