DS$^2$-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning

• URL: http://arxiv.org/abs/2603.12932v2
• Date: Mon, 16 Mar 2026 01:11:56 GMT
• Title: DS$^2$-Instruct: Domain-Specific Data Synthesis for Large Language Models Instruction Tuning
• Abstract summary: DS$2$-Instruct is a framework that generates domain-specific instruction datasets without human supervision.<n>We apply this framework to generate datasets across seven challenging domains, such as mathematics, finance, and logical reasoning.
• Score: 7.350888490884606
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Adapting Large Language Models (LLMs) to specialized domains requires high-quality instruction tuning datasets, which are expensive to create through human annotation. Existing data synthesis methods focus on general-purpose tasks and fail to capture domain-specific terminology and reasoning patterns. To address this, we introduce DS$^2$-Instruct, a zero-shot framework that generates domain-specific instruction datasets without human supervision. Our approach first generates task-informed keywords to ensure comprehensive domain coverage. It then creates diverse instructions by pairing these keywords with different cognitive levels from Bloom's Taxonomy. Finally, it uses self-consistency validation to ensure data quality. We apply this framework to generate datasets across seven challenging domains, such as mathematics, finance, and logical reasoning. Comprehensive evaluation demonstrates that models fine-tuned on our generated data achieve substantial improvements over existing data generation methods.

Related papers

• Domain-Specific Data Synthesis for LLMs via Minimal Sufficient Representation Learning [72.60775633696593]
  Large Language Models can achieve strong performance in specific domains through fine-tuning on domain-specific data.<n>Existing data synthesis approaches rely on explicit domain descriptions expressed in natural language and careful prompt engineering.<n>We propose a novel framework, DOMINO, that learns a minimal sufficient domain representation from reference samples and leverages it to guide the generation of domain-aligned synthetic data.
  arXiv  Detail & Related papers  (2026-05-28T14:57:02Z)
• Generative Data Transformation: From Mixed to Unified Data [57.84692191369066]
  textscTaesar is a emphdata-centric framework for textbftarget-textbfal textbfregeneration.<n>It encodes cross-domain context into target sequences, enabling standard models to learn intricate dependencies without complex fusion architectures.
  arXiv  Detail & Related papers  (2026-02-26T08:30:09Z)
• Scaling Towards the Information Boundary of Instruction Set: InfinityInstruct-Subject Technical Report [11.70656700216213]
  Construction of high-quality instruction datasets is crucial for enhancing model performance and generalizability.<n>We propose a systematic instruction data synthesis framework, which integrates a hierarchical labeling system, an informative seed selection algorithm, and a model deficiency diagnosis.<n>Based on this framework, we construct InfinityInstruct-Subject, a high-quality dataset containing 1.5 million instructions.
  arXiv  Detail & Related papers  (2025-07-09T15:59:02Z)
• Organize the Web: Constructing Domains Enhances Pre-Training Data Curation [129.27104172458363]
  We develop a framework for organizing web pages in terms of both their topic and format.<n>We automatically annotate pre-training data by distilling annotations from a large language model into efficient curations.<n>Our work demonstrates that constructing and mixing domains provides a valuable complement to quality-based data curation methods.
  arXiv  Detail & Related papers  (2025-02-14T18:02:37Z)
• $\textbf{Only-IF}$:Revealing the Decisive Effect of Instruction Diversity on Generalization [1.6958018695660049]
  We show that $textbfonly emerges$ when training data is diversified enough across semantic domains. We extend our analysis to real-world scenarios, including fine-tuning of $textit$textbfspecialist$$ and $textit$textbfgeneralist$$ models.
  arXiv  Detail & Related papers  (2024-10-07T03:15:11Z)
• A Unified Data Augmentation Framework for Low-Resource Multi-Domain Dialogue Generation [52.0964459842176]
  Current state-of-the-art dialogue systems heavily rely on extensive training datasets. We propose a novel data textbfAugmentation framework for textbfMulti-textbfDomain textbfDialogue textbfGeneration, referred to as textbfAMD$2$G. The AMD$2$G framework consists of a data augmentation process and a two-stage training approach: domain-agnostic training and domain adaptation training.
  arXiv  Detail & Related papers  (2024-06-14T09:52:27Z)
• Dial-insight: Fine-tuning Large Language Models with High-Quality Domain-Specific Data Preventing Capability Collapse [4.98050508891467]
  We propose a two-stage approach for the construction of production prompts designed to yield high-quality data. This method involves the generation of a diverse array of prompts that encompass a broad spectrum of tasks and exhibit a rich variety of expressions. We introduce a cost-effective, multi-dimensional quality assessment framework to ensure the integrity of the generated labeling data.
  arXiv  Detail & Related papers  (2024-03-14T08:27:32Z)
• Unsupervised Domain Adaptive Learning via Synthetic Data for Person Re-identification [101.1886788396803]
  Person re-identification (re-ID) has gained more and more attention due to its widespread applications in video surveillance. Unfortunately, the mainstream deep learning methods still need a large quantity of labeled data to train models. In this paper, we develop a data collector to automatically generate synthetic re-ID samples in a computer game, and construct a data labeler to simultaneously annotate them.
  arXiv  Detail & Related papers  (2021-09-12T15:51:41Z)
• Learning to Synthesize Data for Semantic Parsing [57.190817162674875]
  We propose a generative model which models the composition of programs and maps a program to an utterance. Due to the simplicity of PCFG and pre-trained BART, our generative model can be efficiently learned from existing data at hand. We evaluate our method in both in-domain and out-of-domain settings of text-to-Query parsing on the standard benchmarks of GeoQuery and Spider.
  arXiv  Detail & Related papers  (2021-04-12T21:24:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.