Summary of Tqa-bench: Evaluating Llms For Multi-table Question Answering with Scalable Context and Symbolic Extension, by Zipeng Qiu et al.
TQA-Bench: Evaluating LLMs for Multi-Table Question Answering with Scalable Context and Symbolic Extension
by Zipeng Qiu, You Peng, Guangxin He, Binhang Yuan, Chen Wang
First submitted to arxiv on: 29 Nov 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); Information Retrieval (cs.IR)
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 paper presents TQA-Bench, a new benchmark designed to evaluate large language models (LLMs) in question answering (QA) tasks over complex multi-table relational data. The existing benchmarks primarily focus on single-table QA, which fails to capture the intricacies of reasoning across multiple tables. The proposed benchmark incorporates diverse relational database instances from real-world public datasets and introduces a flexible sampling mechanism to create tasks with varying multi-table context lengths. The evaluation framework integrates symbolic extensions to assess LLMs’ reasoning capabilities beyond simple data retrieval or probabilistic pattern matching. A range of LLMs, from 7 billion to 70 billion parameters, are evaluated, revealing critical insights into their performance in multi-table QA. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper creates a new benchmark for evaluating large language models (LLMs) on complex question answering tasks over relational data. This is important because current benchmarks only test single-table questions, which isn’t like real-life where we often have to find answers across multiple tables. The new benchmark uses real-world datasets and lets users create different lengths of multi-table context. It also helps us see how well the models can reason and think beyond just finding data. We tested many LLMs with different numbers of parameters and learned some important things about their abilities. |
Keywords
» Artificial intelligence » Pattern matching » Question answering
