跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 2026-10-15AI 評分47

Data World: Scaling Data Environments from Scratch

Data World: Scaling Data Environments from Scratch

AI 導讀

Data World 是一個多智能體引擎,從明確機制合成數據環境、任務及評分規準,用於訓練模型發掘分析洞見,毋須軌跡蒸餾。DataWorld-Bench 涵蓋 50 個未參與訓練的世界及 16 個領域;以合成任務進行 RL-zero 訓練後,DataWorld-35B-A3B 在四項基準的分數均高於 Qwen3.6-35B-A3B,提升 7.06 至 20.90 分。

正文

Published on Oct 15

Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

A multi-agent engine synthesizes data environments and tasks to train models for analytical insight discovery, improving performance across benchmarks without trajectory distillation.

Data agents use code to process heterogeneous files and derive analytical insights. Scaling their training faces costly environment curation, privacy and intellectual-property restrictions, and difficulty verifying open-ended findings without known generative truth. Our core insight is to specify the world behind the files before generating the files themselves. We introduce Data World, a multi-agent engine that synthesizes environments, tasks, and rubrics from explicit mechanisms. Independent audits check target recoverability from observed files; rubrics support evaluation and RL rewards. Human preference-based curation for question realism selects 100 tasks across 50 held-out worlds and 16 domains for DataWorld-Bench. The engine also yields 2,135 training tasks from 1,037 disjoint worlds. Evaluation of ten frontier models shows that the benchmark remains challenging, particularly for insight discovery, with a consistent computation-insight gap. RL-zero training on the synthesized tasks produces DataWorld-35B-A3B from Qwen3.6-35B-A3B without trajectory distillation. Under matched evaluation, it improves over its base by 7.06 points on DataWorld-Bench, 14.28 on DAComp, 20.90 on DABstep-Hard, and 12.13 on JobBench, a broader occupational benchmark.

View arXiv page View PDF Project page GitHub Add to collection

Get this paper in your agent:

hf papers read 2610.21656

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 1

Zhanli-Li/DataWorld-35B-A3B

Reinforcement Learning •

35B •

Updated about 1 hour ago

•

28

•

1

Datasets citing this paper 3

Zhanli-Li/DataWorld-Bench

Updated about 1 hour ago

•

113

•

1

Zhanli-Li/DataWorld-Train-2k-part1

RL Environment •

Updated 31 minutes ago

•

3

•

1

Zhanli-Li/DataWorld-Train-2k-part2

RL Environment •

Updated 1 minute ago

•

1

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2610.21656 in a Space README.md to link it from this page.

Collections including this paper 1

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co