# Explainable Moral Values: a neuro-symbolic approach to value classification

arXiv: https://arxiv.org/abs/2410.12631 · HTML: https://mlsanity.com/paper/2410.12631/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 29 references that are in the map.
- First posted on arXiv 16 Oct 2024.

## Builds on (most cited first, 9 of 9)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing](https://mlsanity.com/paper/2107.13586/) (2021)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [Parametric UMAP embeddings for representation and semi-supervised learning](https://mlsanity.com/paper/2009.12981/) (2020)
- [Pragmatic-Pedagogic Value Alignment](https://mlsanity.com/paper/1707.06354/) (2017)
- [The Touch'e23-ValueEval Dataset for Identifying Human Values behind Arguments](https://mlsanity.com/paper/2301.13771/) (2023)
- [The Moral Foundations Reddit Corpus](https://mlsanity.com/paper/2208.05545/) (2022)
- [Value alignment: a formal approach](https://mlsanity.com/paper/2110.09240/) (2021)
- [Sandra -- A Neuro-Symbolic Reasoner Based On Descriptions And Situations](https://mlsanity.com/paper/2402.00591/) (2024)

## Built on by (most cited first, 1 of 1)

- [Learning the Value Systems of Agents with Preference-based and Inverse Reinforcement Learning](https://mlsanity.com/paper/2602.04518/) (2026)

## Abstract (arXiv)

This work explores the integration of ontology-based reasoning and Machine Learning techniques for explainable value classification. By relying on an ontological formalization of moral values as in the Moral Foundations Theory, relying on the DnS Ontology Design Pattern, the textitsandra neuro-symbolic reasoner is used to infer values (fomalized as descriptions) that are satisfied by a certain sentence. Sentences, alongside their structured representation, are automatically generated using an open-source Large Language Model. The inferred descriptions are used to automatically detect the value associated with a sentence. We show that only relying on the reasoner's inference results in explainable classification comparable to other more complex approaches. We show that combining the reasoner's inferences with distributional semantics methods largely outperforms all the baselines, including complex models based on neural network architectures. Finally, we build a visualization tool to explore the potential of theory-based values classification, which is publicly available at http://xmv.geomeaning.com/.