LG AI Research 发布 EXAONE Tabular：重新发掘字段数据的价值
Rediscovering the Value of Field Data: ‘EXAONE Tabular’ - LG AI Research BLOG
LG AI Research 发布 EXAONE Tabular 1.0，用于表格数据的分类与回归任务，论文见 arXiv:2608.25774。该模型采用 CAST 跨轴注意力架构，让列摘要与行摘要仅通过单元格表示交换信息，并配备分类与回归专用任务头，支持以支持集和查询集进行上下文学习推理。
正文
当前显示原文，中文全文译文尚未就绪。
Moonjung Eo
2026.09.09
Figure 1. Accuracy (Elo) vs. Inference Latency Pareto Frontier (Left: Classification, Right: Regression).
Figure 2. Effects of Tuning (Orange) and Ensembling (Green) vs. Default Settings (Left: Classification, Right: Regression).
Figure 3. (1) SCM-Based Synthetic Data Generation: Generates training tasks by combining causal graphs, functional mechanisms, and missingness/noise patterns.
(2) CAST’s Cross-Axis Attention Architecture: Column summaries ($S_c$) and row summaries ($S_r$) exchange information exclusively through cell representations.
(3) Task-Specific Heads: Performs task-tailored training using dedicated heads for classification and regression, respectively.
(4) In-Context Learning Inference: Takes a support set and a query set as inputs to output class probabilities or predictive distributions. (Source: EXAONE Tabular 1.0 (http://arxiv.org/abs/2608.25774))