From Table to Cell: Attention for Better Reasoning with TABALIGN
Tung Sum Thomas Kwok, Zeyong Zhang, Xinyu Wang, Chunhe Wang, Xiaofeng Lin, Hanwei Wu, Lei Ding, Guang Cheng, Zhijiang Guo

TL;DR
This paper introduces TABALIGN, a new framework for multi-step reasoning over tables using diffusion language models and cell attention verification, significantly improving accuracy and efficiency.
Contribution
It proposes a planned reasoning framework with a bidirectional DLM planner and a cell attention verifier, enhancing permutation invariance and reasoning accuracy.
Findings
DLMs produce more human-aligned, permutation-stable cell attention than autoregressive models.
TABALIGN improves accuracy by 15.76 percentage points over strong open-source baselines.
Cleaner DLM plans accelerate reasoning execution by 44.64%.
Abstract
Multi-step LLM reasoning over structured tables fails because planning and execution share no explicit cell-grounding contract. Existing methods constrain the planner to a left-to-right factorization at odds with table permutation invariance, and score intermediate states by generated content alone, overlooking cell grounding. We conduct a pilot study showing that diffusion language models (DLMs) produce more human-aligned and permutation-stable cell attention on tables than autoregressive models, with a 40.2% median reduction in attention-AUROC variability under row reordering. Motivated by this, we propose TABALIGN, a planned table reasoning framework that operationalizes the contract. TABALIGN pairs a masked DLM planner, whose bidirectional denoising emits plan steps as binary cell masks, with TABATTN, a lightweight verifier trained on 1,600 human-verified attention standards to…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
