Publications

Argument schemes are structured templates that capture common patterns of reasoning. We explore whether schemes can guide the automatic construction of formal proofs that derive an argument’s claim from its premises. To do this, we propose a pipeline that translates natural language into proofs in Isabelle/HOL, using argument schemes. The schemes are formalised systematically as inference rules in higher-order logic, and are used to structure proof generation within an iterative LLM-driven reasoning loop. Evaluating on a dataset of arguments that fully instantiate their schemes across 18 scheme types, we compare this scheme-guided approach against a baseline in which no scheme is provided. There are two sources of incompatibility between schemes and natural language arguments: one where the scheme’s conclusion does not directly match the argument’s claim, and one where the natural language premises are insufficient to satisfy the scheme’s requirements. We find that scheme guidance only marginally improves results overall, though when the scheme’s requirements are satisfied, the pipeline reliably closes the final inferential step; the bottleneck is that natural language arguments rarely make the required information explicit, and LLMs struggle to supply the necessary connections. Our error analysis reveals four key factors contributing to this incompatibility, and we suggest future directions to address them.

@incollection{saadatyazdi2026argumentation,
  title = {Do Argumentation Schemes Help? An Assessment of Their Role in Formalising Natural Language Arguments},
  author = {Ameer Saadat-Yazdi and Víctor David and Anthony Hunter and Serena Villata},
  booktitle = {Computational Models of Argument (COMMA 2026)},
  volume = {428},
  pages = {329--340},
  year = {2026},
  doi = {10.3233/faia260823},
  url = {https://doi.org/10.3233/faia260823}
}

One of the key distinctions between classical formal logic and argumentative discourse is that the latter is a social process, often taking place through the exchange of language. The nature of language and social discourse means that arguments often employ unspoken inferential rules that are shaped by norms. While this process is informal, in the sense that these inferential rules and the procedure for exchanging arguments are very rarely explicitly established, patterns can be seen to emerge. This thesis seeks to model how implicit inferences arise in natural language arguments, using methods in computational linguistics to make these inferences explicit. The work described here takes an interdisciplinary approach using computational methods to operationalise existing theories, while also ensuring that computational methods are grounded in theoretical frameworks. This allows a theory of argumentation to be developed that is faithful to real-world data, while constraining the computational models to provide a meaningful and interpretable analysis of arguments. This thesis contributes to the existing literature by presenting various technical approaches to integrating world-knowledge in argument analysis; extending an existing theory of argumentation with an account of how seemingly non-sequitur arguments can be analysed; outlining how data can be reliably annotated for explicating implicit argumentative material; and, as a culmination of the results of this thesis, presents a method for identifying implicit premises that are crucial for generating critical questions to assess implicit assumptions behind arguments.

@phdthesis{saadatyazdi2026invisible,
  title = {Invisible threads: computational insights into the hidden rules of argumentation},
  author = {Ameer Saadat-Yazdi},
  school = {University of Edinburgh},
  year = {2026},
  doi = {10.7488/era/7016},
  url = {https://doi.org/10.7488/era/7016}
}

Abstract Enthymemes are arguments that are not fully articulated, often omitting a connection between premise and conclusion but sometimes also other information that is crucial for their interpretation. This implicitness poses challenges for the analysis and evaluation of argumentative discourse. We use the concept of “argument form” as employed in the argument classification framework of the Periodic Table of Arguments to address this issue. By developing an algorithmic procedure grounded in this concept, we provide a method for explicating missing statements and connections condensed in enthymemes. Our approach contributes to understanding the pragmatics of argumentation, as it offers a formal framework for analysing how the interpretation of implicit elements in argumentation arises from apparent non-sequiturs. The algorithmic procedure we developed can function as a guideline for human annotation of argumentative discourse and is also suitable for implementation in (AI-assisted) annotation software for argument mining.

@article{saadatyazdi2025finding,
  title = {Finding the Missing Link: An Algorithmic Approach to Reconstructing Enthymemes},
  author = {Ameer Saadat-Yazdi and Jean H. M. Wagemans},
  journal = {Argumentation (journal)},
  year = {2025},
  doi = {10.1007/s10503-025-09682-z},
  url = {https://doi.org/10.1007/s10503-025-09682-z}
}

Critical questions are essential resources to provoke critical thinking when encountering an argumentative text.We present our system for the Critical Questions Generation (CQs-Gen) Shared Task at ArgMining 2025.Our approach leverages large language models (LLMs) with chain-of-thought prompting to generate critical questions guided by Walton's argumentation schemes.For each input intervention, we conversationally prompt LLMs to instantiate the corresponding argument scheme template to first obtain structured arguments, and then generate relevant critical questions.Following this, we rank all the available critical questions by prompting LLMs to select the top 3 most helpful questions based on the original intervention text.This combination of structured argumentation theory and step-by-step reasoning enables the generation of contextually relevant and diverse critical questions.Our pipeline achieves competitive performance in the final test set, showing its potential to foster critical thinking given argumentative text and detect missing or uninformed claims. .

@inproceedings{zhou2025daydreamer,
  title = {DayDreamer at CQs-Gen 2025: Generating Critical Questions through Argument Scheme Completion},
  author = {Wei Guo Zhou and Ameer Saadat-Yazdi and Nadin Kökciyan},
  booktitle = {12th Workshop on Argument Mining (ArgMining 2025)},
  pages = {289--294},
  year = {2025},
  doi = {10.18653/v1/2025.argmining-1.27},
  url = {https://doi.org/10.18653/v1/2025.argmining-1.27}
}

Reviews are valuable resources for customers making purchase decisions in online shopping.However, it is impractical for customers to go over the vast number of reviews and manually conclude the prominent opinions, which prompts the need for automated opinion summarization systems.Previous approaches, either extractive or abstractive, face challenges in automatically producing grounded aspectcentric summaries.In this paper, we propose a novel summarization system that not only captures predominant opinions from an aspect perspective with supporting evidence, but also adapts to varying domains without relying on a pre-defined set of aspects.Our proposed framework, ASESUM, summarizes viewpoints relevant to the critical aspects of a product by extracting aspect-centric arguments and measuring their salience and validity.We conduct experiments on a real-world dataset to demonstrate the superiority of our approach in capturing diverse perspectives of the original reviews compared to new and existing methods.

@inproceedings{zhou2025aspect,
  title = {Aspect-Based Opinion Summarization with Argumentation Schemes},
  author = {Wendi Zhou and Ameer Saadat-Yazdi and Nadin Kökciyan},
  booktitle = {12th Workshop on Argument Mining (ArgMining 2025)},
  pages = {116--125},
  year = {2025},
  doi = {10.18653/v1/2025.argmining-1.11},
  url = {https://doi.org/10.18653/v1/2025.argmining-1.11}
}

In argumentation theory, argument schemes are a characterisation of stereotypical patterns of inference.There has been little work done to develop computational approaches to identify these schemes in natural language.Moreover, advancements in recognizing textual entailment lack a standardized definition of inference, which makes it challenging to compare methods trained on different datasets and rely on the generalisability of their results.In this work, we propose a rigorous approach to align entailment recognition with argumentation theory.Wagemans' Periodic Table of Arguments (PTA), a taxonomy of argument schemes, provides the appropriate framework to unify these two fields.To operationalise the theoretical model, we introduce a tool to assist humans in annotating arguments according to the PTA.Beyond providing insights into non-expert annotator training, we present Kialo-PTA24, the first multi-topic dataset for the PTA.Finally, we benchmark the performance of pre-trained language models on various aspects of argument analysis.Our experiments show that the task of argument canonicalisation poses a significant challenge for state-of-the-art models, suggesting an inability to represent argumentative reasoning and a direction for future investigation.1. We conduct an annotation study to rephrase natural language arguments into structured templates and provide insights into how to train non-expert annotators to perform this analysis.2. We introduce ArgNotator, a tool that assists humans in annotating arguments according to the PTA.3. We construct Kialo-PTA24 -the first multi-topic dataset of argument types annotated according to the PTA.4. We compare the performance of state-of-theart models for two annotation subtasks.For the substance classification task, we benchmark the performance of a number of BERT-based models.For the argument canonicalisation task, we evaluate the performance of two large language models (FLAN-T5, LLAMA2) in both pre-trained and few-shot settings.The dataset, experimental setup, annotation tool and training materials can all be found on GitLab. 1

@inproceedings{saadatyazdi2024beyond,
  title = {Beyond Recognising Entailment: Formalising Natural Language Inference from an Argumentative Perspective},
  author = {Ameer Saadat-Yazdi and Nadin Kökciyan},
  booktitle = {62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024)},
  pages = {9620--9636},
  year = {2024},
  doi = {10.18653/v1/2024.acl-long.520},
  url = {https://doi.org/10.18653/v1/2024.acl-long.520}
}

Individuals engage in arguments on an everyday basis as they seek to obtain information about current affairs and engage with social media. While fact-checkers are available to help dispel misinformation, it is almost impossible for users to verify every single claim they encounter. This means that oftentimes, it is left to the user to decide whether a claim is well supported. To address this, we have developed a Web interface that allows users to input an argument, and our developed framework automatically detects its validity (soundness of logical deduction) and novelty (whether the argument is non-circular). Our Web-based tool could be used by social media users who wish to evaluate the information they consume. As part of one of the modules developed at the University of Edinburgh, our tool will be deployed as a teaching tool for the students who study argumentation.

@inproceedings{chausson2023based,
  title = {A Web-based Tool for Detecting Argument Validity and Novelty},
  author = {Sandrine Chausson and Ameer Saadat-Yazdi and Xue Li and Jeff Z. Pan and Vaishak Belle and Nadin Kökciyan and Björn Roß},
  booktitle = {22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2023), Demo Track},
  pages = {3053--3055},
  year = {2023},
  doi = {10.65109/rags2524},
  url = {https://doi.org/10.65109/rags2524}
}

Argument mining seeks to extract arguments and their structure from unstructured texts.Identifying relations (such as attack, support, and neutral) between argumentative units is a challenging task because two units may be related to each other via implicit inferences.These inferences often rely on external commonsense knowledge to discover how one argumentative unit relates to another.State-of-the-art methods, however, rely on predefined knowledge graphs, and thus might not cover target pairs of argumentative units well.We introduce a new generative approach to finding inference chains that connect these pairs by making use of the Commonsense Transformer (COMET).We evaluate our approach on three datasets for both the two-label (attack/support) and three-label (attack/support/neutral) tasks.Our approach significantly outperforms the state-of-the-art, by 2-5% in F1 score, on two out of the three datasets with minor improvements on the remaining one.

@inproceedings{saadatyazdi2023uncovering,
  title = {Uncovering Implicit Inferences for Improved Relational Argument Mining},
  author = {Ameer Saadat-Yazdi and Jeff Z. Pan and Nadin Kökciyan},
  booktitle = {17th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2023)},
  pages = {2484--2495},
  year = {2023},
  doi = {10.18653/v1/2023.eacl-main.182},
  url = {https://doi.org/10.18653/v1/2023.eacl-main.182}
}

The ArgMining 2022 Shared Task is concerned with predicting the validity and novelty of an inference for a given premise and conclusion pair. We propose two feed-forward network based models (KEViN1 and KEViN2), which combine features generated from several pretrained transformers and the WikiData knowledge graph. The transformers are used to predict entailment and semantic similarity, while WikiData is used to provide a semantic measure between concepts in the premise-conclusion pair. Our proposed models show significant improvement over RoBERTa, with KEViN1 outperforming KEViN2 and obtaining second rank on both subtasks (A and B) of the ArgMining 2022 Shared Task.

@inproceedings{saadatyazdi2022kevin,
  title = {KEViN: A Knowledge Enhanced Validity and Novelty Classifier for Arguments},
  author = {Ameer Saadat-Yazdi and Xue Li and Sandrine Chausson and Vaishak Belle and Björn Ross and Jeff Z. Pan and Nadin Kökciyan},
  booktitle = {9th Workshop on Argument Mining (ArgMining 2022)},
  pages = {104--110},
  year = {2022},
  url = {https://aclanthology.org/2022.argmining-1.9}
}

Alzheimer's disease is a debilitating disease in the elderly, and is an increasing burden to the society due to an aging population. In this paper, we apply topological data analysis to structural MRI scans of the brain, and show that topological invariants make accurate predictors for Alzheimer's. Using the construct of Betti Curves, we first show that topology is a good predictor of Age. Then we develop an approach to factor out the topological signature of age from Betti curves, and thus obtain accurate detection of Alzheimer's disease. Experimental results show that topological features used with standard classifiers perform comparably to recently developed convolutional neural networks. These results imply that topology is a major aspect of structural changes due to aging and Alzheimer's. We expect this relation will generate further insights for both early detection and better understanding of the disease.

@inproceedings{saadatyazdi2021topological,
  title = {Topological Detection of Alzheimer’s Disease Using Betti Curves},
  author = {Ameer Saadat-Yazdi and Rayna Andreeva and Rik Sarkar},
  booktitle = {TDA4MedicalData Workshop @ MICCAI 2021},
  pages = {119--128},
  year = {2021},
  doi = {10.1007/978-3-030-87444-5_12},
  url = {https://doi.org/10.1007/978-3-030-87444-5_12}
}