Summary of Iterative Repair with Weak Verifiers For Few-shot Transfer in Kbqa with Unanswerability, by Riya Sawhney et al.
Iterative Repair with Weak Verifiers for Few-shot Transfer in KBQA with Unanswerability
by Riya Sawhney, Samrat Yadav, Indrajit Bhattacharya, Mausam
First submitted to arxiv on: 20 Jun 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 The proposed model, FUn-FuSIC, tackles the novel task of few-shot transfer for knowledge-based question answering (KBQA) with unanswerable questions. Building upon FuSIC-KBQA, the current state-of-the-art for answerable-only KBQA, FUn-FuSIC introduces Feedback for Unanswerability (FUn), which leverages iterative repair and self-consistency to assess question answerability. The model outperforms existing solutions, including LLM-based and supervised models, on both unanswerable few-shot transfer and answerable few-shot transfer tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper focuses on developing a new method for answering questions that can’t be answered using limited training data. It proposes a solution called FUn-FuSIC, which is better at handling unanswerable questions than other models. The researchers tested their model and found it outperformed others in both answering unanswerable questions and answering answerable questions with only a few examples. |
Keywords
» Artificial intelligence » Few shot » Question answering » Supervised
