TaCube: Pre-computing Data Cubes for Answering Numerical-Reasoning Questions over Tabular Data
Fan Zhou, Mengkang Hu, Haoyu Dong, Zhoujun Cheng, Fan Cheng, Shi Han, Dongmei Zhang
Abstract
Existing auto-regressive pre-trained language models (PLMs) like T5 and BART, have been well applied to table question answering by UNIFIEDSKG and TAPEX, respectively, and demonstrated state-of-the-art results on multiple benchmarks. However, auto-regressive PLMs are challenged by recent emerging numerical reasoning datasets, such as TAT-QA, due to the error-prone implicit calculation. In this paper, we present TaCube, to pre-compute aggregation/arithmetic results for the table in advance, so that they are handy and readily available for PLMs to answer numerical reasoning questions. TaCube systematically and comprehensively covers a collection of computational operations over table segments. By simply concatenating TaCube to the input sequence of PLMs, it shows significant experimental effectiveness. TaCube promotes the F1 score from 49.6% to 66.2% on TAT-QA and achieves new state-of-the-art results on Wik-iTQ (59.6% denotation accuracy). TaCube 's improvements on numerical reasoning cases are even more notable: on TAT-QA, TaCube promotes the exact match accuracy of BARTlarge by 39.6% on sum, 52.5% on average, 36.6% on substraction, and 22.2% on division. We believe that TaCube is a general and portable pre-computation solution that can be potentially integrated to various numerical reasoning frameworks. Data and code will be available at https://github.com/ koalazf99/tacube .
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
5 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- OpenTab: Advancing Large Language Models as Open-domain Table Reasoners cites UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models
- OpenTab: Advancing Large Language Models as Open-domain Table Reasoners cites TAPEX: Table Pre-training via Learning a Neural SQL Executor
- UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models cites TAPEX: Table Pre-training via Learning a Neural SQL Executor
- UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
- QTSumm: Query-Focused Summarization over Tabular Data cites TabFact: A Large-scale Dataset for Table-based Fact Verification
- QTSumm: Query-Focused Summarization over Tabular Data cites TAPEX: Table Pre-training via Learning a Neural SQL Executor
- QTSumm: Query-Focused Summarization over Tabular Data cites MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual Data
- ReAcTable: Enhancing ReAct for Table Question Answering cites TAPEX: Table Pre-training via Learning a Neural SQL Executor
- TAPEX: Table Pre-training via Learning a Neural SQL Executor cites TabFact: A Large-scale Dataset for Table-based Fact Verification
- TAPEX: Table Pre-training via Learning a Neural SQL Executor cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
- TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data cites TabFact: A Large-scale Dataset for Table-based Fact Verification
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples cites TabFact: A Large-scale Dataset for Table-based Fact Verification
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples cites UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples cites TAPEX: Table Pre-training via Learning a Neural SQL Executor
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning Examples cites MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual Data
- MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual Data cites UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Models
Cited by top-tier papers5
- ReAcTable: Enhancing ReAct for Table Question AnsweringYunjia Zhang, Jordan Henkel, Avrilia Floratou, Joyce Cahoon et al.VLDB 2024 · 120 citations
- OpenTab: Advancing Large Language Models as Open-domain Table ReasonersKezhi Kong, Jiani Zhang, Zhengyuan Shen, Balasubramaniam Srinivasan et al.ICLR 2024 · 40 citations
- PACIFIC: Towards Proactive Conversational Question Answering over Tabular and Textual Data in FinanceYang Deng, Wenqiang Lei, Wenxuan Zhang, Wai Lam et al.EMNLP 2022 · 25 citations
- ReasTAP: Injecting Table Reasoning Skills During Pre-training via Synthetic Reasoning ExamplesYilun Zhao, Linyong Nan, Zhenting Qi, Rui Zhang et al.EMNLP 2022 · 15 citations
- QTSumm: Query-Focused Summarization over Tabular DataYilun Zhao, Zhenting Qi, Linyong Nan, Boyu Mi et al.EMNLP 2023 · 4 citations
Builds on11
- TabFact: A Large-scale Dataset for Table-based Fact VerificationWenhu Chen, Hongmin Wang, Jianshu Chen, Yunkai Zhang et al.ICLR 2020 · 674 citations
- TaBERT: Pretraining for Joint Understanding of Textual and Tabular DataPengcheng Yin, Graham Neubig, Wen-tau Yih, Sebastian RiedelACL 2020 · 417 citations
- TAPEX: Table Pre-training via Learning a Neural SQL ExecutorQian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi et al.ICLR 2022 · 347 citations
- UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language ModelsTianbao Xie, Chen Henry Wu, Peng Shi, Ruiqi Zhong et al.EMNLP 2022 · 222 citations
- MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual DataYilun Zhao, Yunxiang Li, Chenying Li, Rui ZhangACL 2022 · 168 citations
Related papers
- TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in FinanceFengbin Zhu, Wenqiang Lei, Youcheng Huang, Chao Wang et al.ACL 2021
- TabDSR: Decompose, Sanitize, and Reason for Complex Numerical Reasoning in Tabular DataChangjiang Jiang, Fengchang Yu, Haihua Chen, Wei Lu et al.EMNLP Findings 2025 · 3 citations
- Injecting Numerical Reasoning Skills into Language ModelsMor Geva, Ankit Gupta, Jonathan BerantACL 2020 · 12 citations
- TAP4LLM: Table Provider on Sampling, Augmenting, and Packing Semi-structured Data for Large Language Model ReasoningYuan Sui, Jiaru Zou, Mengyu Zhou, Xinyi He et al.EMNLP Findings 2024 · 18 citations
- Evaluating LLMs' Mathematical Reasoning in Financial Document Question AnsweringPragya Srivastava, Manuj Malik, Vivek Gupta, Tanuja Ganu et al.ACL Findings 2024 · 27 citations
