Research map / Learning theory and algorithms / Prediction and decision losses

An Extension-Based Argument-Ranking Semantics: Social Rankings in Abstract Argumentation

AAAI 2025 · Learning theory and algorithms › Prediction and decision losses · approach: Ranking models from comparisons · 0 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces a family of argument-ranking semantics obtained by applying social ranking functions (notably lexicographic excellence) to extension rankings, giving a refinement of skeptical/credulous/rejected acceptance, with axiomatic necessary and sufficient conditions.

Problem

Argument-ranking semantics are technically distinct from extension-based acceptance classifications (skeptical, credulous, rejected), so they do not refine these classes. The paper asks how to build rankings that refine skeptical/credulous acceptance.

Approach

Combine extension-ranking semantics (which order sets of arguments by plausibility) with social ranking functions (which rank elements by the quality of sets containing them), generalised to partial orders via a set-rank notion. Arguments are ranked by how often they appear in better-ranked extensions, using lex-cel; axioms (Independence from the worst set, Pareto-efficiency, Dominating set) are s

Key results

• Lex-cel applied to r-co extension rankings yields an argument ranking in which all skeptically accepted arguments precede credulously accepted ones, which precede rejected ones (satisfies sigma-Refinement).
• Independence from the worst set together with Pareto-efficiency implies Dominating set, and together they are sufficient for sigma-C and sigma-sk-C.
• Lex-cel-based semantics satisfy SC, Abs, In, AvsFD, and sigma-C/sigma-sk-C for r-ad, r-co, r-gr, r-pr, r-sst, but violate VP, CP, QP, CT, and others.

Surprising findings

• The lex-cel argument ranking violates void precedence (VP): an attacked argument can be ranked above an unattacked one.
• The ceteris paribus majority and ordinal Banzhaf social-ranking-based semantics violate self-contradiction (SC), so they are unsuitable for argumentation.

Limitations

• Many principle compliance results are stated only in the supplementary material or left for future work (full lex-cel principle analysis deferred).
• The paper is purely theoretical with no empirical evaluation or implementation.

Experimental setup

Baselines: singleton argument-ranking semantics serialisability-based argument-ranking semantics ordinal Banzhaf-based argument ranking ceteris paribus majority-based argument ranking

Builds on: extension-ranking semantics lexicographic excellence operator argument-ranking semantics principles

Closest papers by meaning

• A Comparative Study of Ranking-Based Semantics for Abstract Argumentation (AAAI 2016 · similarity 0.93)
• Ranking Sets of Defeasible Elements in Preferential Approaches to Structured Argumentation: Postulates, Relations, and Characterizations (AAAI 2021 · similarity 0.89)
• Ranking-Based Semantics for Sets of Attacking Arguments (AAAI 2020 · similarity 0.89)
• Strong Explanations in Abstract Argumentation (AAAI 2021 · similarity 0.87)
• Bilateral Gradual Semantics for Weighted Argumentation (AAAI 2024 · similarity 0.86)
• Redefining ABA+ Semantics via Abstract Set-to-Set Attacks (AAAI 2024 · similarity 0.84)