Selective Rationalization | Language Technologies Lab

Knowledge Extraction from Rationalization

Mon, 02 Mar 2026 00:00:00 +0000

Description:
Rationalization is a type of example-specific explanation. However, samples belonging to the same class might share similar rationales. The idea is to define ways to go from a local explanation (i.e., rationalization) to a global explanation (i.e., knowledge base) by aggregating and summarizing extracted rationales. This can be done with LLMs (e.g., prompting techniques) or other solutions.

Contact: Federico Ruggeri

References:

A Game Theoretic Approach to Class-wise Selective Rationalization
Shiyu Chang, Yang Zhang, Mo Yu, Tommi S. Jaakkola. 33rd Conference on Neural Information Processing Systems (NeurIPS), Vancouver, Canada, 2019.
PDF

Mixture of Experts for Rationalization

Mon, 02 Mar 2026 00:00:00 +0000

Description:
Mixture of Experts (MoE) is a technique whereby several models are trained on the same data, each specializing in a certain subset. MoE have been shown to be successful in a variety of applications and their original formulation dates back early 2000s. The idea is to understand whether we can develop a MoE model for selective rationalization to address interlocking.

Contact: Federico Ruggeri

References:

A Survey on Mixture of Experts in Large Language Models
W. Cai, J. Jiang, F. Wang, J. Tang, S. Kim and J. Huang.
In IEEE Transactions on Knowledge and Data Engineering, vol. 37, no. 7, pp. 3896-3915, July 2025.
DOI

Rationalization via LLMs

Mon, 02 Mar 2026 00:00:00 +0000

Description:
LLMs are ubiquitous in NLP. Our aim is to evaluate LLM capabilities in performing selective rationalization via prompting. How do they compare w.r.t. traditional SPP models?

Contact: Federico Ruggeri

References:

Towards Faithful Explanations: Boosting Rationalization with Shortcuts Discovery
Linan Yue, Qi Liu, Yichao Du, Li Wang, Weibo Gao, Yanqing An.
The Twelfth International Conference on Learning Representations, 2024.
PDF

Learning Robust Rationales for Model Explainability: A Guidance-Based Approach
S Hu, K Yu.
Proceedings of the AAAI Conference on Artificial Intelligence, 2024.
DOI | PDF

Structured Rationalization via Tree kernel methods

Mon, 02 Mar 2026 00:00:00 +0000

Description:
There are several techniques for transforming text into abstract structured content (AMR graphs, Parse trees, etc…). We are interested in applying rationalization in these contexts by also enforcing some structural constraints depending on the given scenario of application. The constraints describe which type of allowed structured the rationalization system can extract. In the case of tree kernels, these structures are different types of trees.

Contact: Federico Ruggeri

References:

Tree-constrained Graph Neural Networks for Argument Mining
Federico Ruggeri, Marco Lippi, Paolo Torroni
September 2021
PDF