Summary of Syntaxshap: Syntax-aware Explainability Method For Text Generation, by Kenza Amara et al.
SyntaxShap: Syntax-aware Explainability Method for Text Generation
by Kenza Amara, Rita Sevastjanova, Mennatallah El-Assady
First submitted to arxiv on: 14 Feb 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel approach to explaining sequence-to-sequence tasks using textual data is presented in this paper, focusing on ensuring the explainability of large language models in safety-critical domains. The proposed method, SyntaxShap, takes into account syntax-based syntactic dependencies and extends Shapley values for parsing-based constraints. This model-agnostic approach evaluates the faithfulness, coherency, and semantic alignment of explanations with autoregressive models. Compared to state-of-the-art methods, SyntaxShap shows promise in producing more faithful and coherent explanations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large language models are becoming increasingly important in safety-critical domains, but we need to make sure they’re transparent about their predictions. This paper looks at a new way to explain how these models work when generating text. The method, called SyntaxShap, considers the structure of the words and sentences being generated. It’s compared to other methods that try to explain text generation tasks, showing that it does a better job of providing accurate and clear explanations. |
Keywords
» Artificial intelligence » Alignment » Autoregressive » Parsing » Syntax » Text generation
