data-curation

The author argues that RL environments serve as the essential data for building AI agents, enabling systematic training, prompt optimization, and evaluation rather than manual iteration.

Introduces Poplar, a scalable Specify-Render-Inspect pipeline for synthesizing human-centric image datasets, and releases Poplar-9K, a curated dataset of 9,401 image-text pairs with auditable inspection records.

DecoupleMix introduces a systematic framework for optimizing pretraining data mixtures for Vision-Language Models by decoupling inter-class and intra-class ratio search, using convex optimization to improve scalability and performance over heuristic baselines.

This paper presents an engineering study adapting NVIDIA Nemotron 3.5 ASR Streaming 0.6B to Kikuyu, Dholuo, and Kalenjin, achieving 42.97% and 33.98% WER on internal sets for Kikuyu and Dholuo, respectively, through data-centric techniques including corpus auditing, normalization, and streaming evaluation.

This paper introduces BatteryLake, a governed data lakehouse that uses LLM agents for evidence-grounded metadata extraction and schema mapping, with human-in-the-loop verification, to curate heterogeneous battery aging datasets and release an open benchmark.

This paper presents methodological contributions for physics-informed machine learning under small-data constraints, using an abrasive waterjet milling dataset of 155 points. It shows that data curation choices, evaluation design, and physics integration form matter significantly, with Gaussian Process variants outperforming other models.

NVIDIA discusses the importance of open and synthetic data for building robust AI agents, highlighting their Nemotron open datasets for training, reasoning, and tool-use.

CurateEvo is a failure-driven dynamic evolution framework for agentic post-training data curation. It iteratively rewrites curation strategies using failed trajectories, improving effectiveness and efficiency on benchmarks like ACEBench-Agent and BFCL-V4.

MedPMC is an automated framework that transforms medical literature into high-fidelity multimodal data for foundation models, achieving significant improvements across multiple benchmarks and clinical settings.

This paper proposes a generator-agnostic post-generation curation method that selects informative subsets of synthetic images by splitting real classes into canonical homogeneous and non-redundant heterogeneous subsets, and scoring synthetic images via a fidelity-diversity criterion. It consistently outperforms existing data-selection baselines and matches real-data performance with up to 40% fewer synthetic samples.

This paper introduces DataComp-VLM (DCVLM), a comprehensive benchmark for evaluating data curation strategies for vision-language models. The authors find that data mixing, rather than filtering, significantly improves performance, and their resulting DCVLM-Baseline dataset achieves state-of-the-art results on 33 downstream tasks.

The article explains how GEPA (Genetic-Pareto Optimization) within DSPy is used for efficient prompt tuning, specifically applied to pretraining data curation at Microsoft AI, allowing researchers to replace manual prompt engineering with automated compute-driven optimization.

This paper introduces OpenThoughts-Agent, an open-source data curation pipeline for training agentic language models, achieving a 44.8% average accuracy across seven benchmarks and outperforming prior open datasets through systematic experiments.

DRIFT proposes a method that uses on-policy influence functions to refine training data distribution for supervised fine-tuning of large language models, consistently improving performance ceilings over existing baselines.

A fine-grained study of narrative features in web-scale LLM pretraining data, introducing NarraBERT and NarraDolma to measure narrative patterns and their distribution across sources.

This paper presents STG, a structured testbench generation framework for LLM-driven hardware design workflows that reduces token cost and improves verification reliability compared to existing prompt-based approaches.

Researchers introduce Curation-Bench, a benchmark to evaluate whether generalist coding agents can automate the iterative data curation loop in AI development. Results show agents can match strong baselines within ten iterations, but reliable data research requires scaffolded method adaptation rather than open-ended prompting alone.

This paper explores whether generalist coding agents (Claude Code, Codex, etc.) can automate data curation loops, achieving published baselines within 10 iterations but revealing a gap in exploring new methods. A scaffold that forces agents to adapt prior research yields policies that beat baselines using 10x less data.

This paper introduces Autonomous Agentic Data Engineering, a task where LLMs autonomously execute end-to-end data curation pipelines for model specialization, showing significant performance gains (e.g., GPT-5.2 improves a student model by 57.29%).

LoMo proposes a data curation method that reformulates single-modality prompts into interleaved multimodal sequences to improve cross-modal representation alignment in vision-language models, achieving consistent gains on multiple benchmarks.