data-generation

This paper introduces a two-level framework for agentic data generation using the ACE lens, focusing on accuracy, complexity, and diversity to create effective training data for LLM agents.

This paper proposes a circuit-grounded framework that leverages mechanistic interpretability for controllable data generation in language models, introducing SAMS for stage-aware data scheduling to improve training performance.

HOTFIXR is a data generation framework that targets multilingual reasoning weaknesses in LLMs to improve cross-lingual performance using synthetic data without sacrificing overall capability.

Data Turnstile is an open-source framework for generating high-quality synthetic function-calling training data from API specifications. Fine-tuning small language models with this data significantly improves their tool-use performance, closing the gap with much larger models.

Introduces SKT, a verified data synthesis pipeline for skill-use training of language model agents, producing 4,000 task packages and 27,164 verified trajectories from 2,000 public skills. Supervised fine-tuning on SKT-generated trajectories consistently improves skill-use performance across models and agent harnesses.

Introduces Pegasus, a low-resource framework that translates human demonstration videos into robot-executable data using graph-based task representation, hierarchical affordance latent space, and closed-loop physics verification, aiming to turn hardware data collection into scalable knowledge transfer.

The paper proposes EmoTrace, a multi-turn dialogue generation framework for psychological support that models seekers' emotional trajectories to improve empathy and emotional richness in counselor responses, outperforming existing methods.

HiFi-UMI introduces a portable data-production system for robot-free UMI data that achieves high trajectory accuracy using stereo-inertial SLAM and wide-angle cameras. Training manipulation policies on this data alone enables zero-shot deployment on real robots, matching or exceeding teleoperation baselines across several model families, and the authors open-source a 2,000-hour high-fidelity dataset.

G-MAD is an open-source framework using Arma 3 to generate synchronized multi-view RGB-T data for aerial object detection, addressing limitations of real-world datasets. It also introduces the AMOD benchmark.

FlowMimic presents a method for mask-free visual editing and generation across video and image modalities using pixel-pair warped flow fields, enabling real-time video editing data generation from image editing samples and aligning modality capabilities through mimicry losses.

This paper proposes an environment-free synthetic data generation approach for training API-calling LLM agents, using LLMs as on-the-fly digital world models to generate trajectories, and shows significant performance gains on challenging benchmarks.

A tweet thread discussing the iterative process of building evaluations and environments for AI agents, emphasizing human-agent collaboration and the importance of data and verifier design.

This research examines whether LLMs can generate synthetic consumer data for projective techniques, comparing human and LLM responses on city tourism perceptions and finding substantial overlap but differences in style and diversity.

This paper introduces Autodata, a method that uses an agentic 'data scientist' AI to automate the creation of high-quality synthetic datasets through iterative generation, verification, and refinement, specifically optimized for reinforcement learning (GRPO) to improve reasoning in language models.

TDGT is a web-based toolkit for synthetic tabular data generation that introduces the Adaptive Bayesian Mixture Synthesizer (ABMS) algorithm and a hybrid VAE-ABMS model, with GPU acceleration and comprehensive fidelity assessment.

Autodata introduces an agentic data scientist that iteratively generates and refines synthetic training data, with meta-optimization to further improve data quality, achieving better results on computer science and legal reasoning tasks.

Meta's new paper 'Autodata' introduces an agentic data scientist that generates and meta-optimizes synthetic training data, significantly outperforming standard methods and enabling a small 4B model to beat a 397B baseline in legal tasks.

Meta releases Autodata, an agentic data scientist that generates high-quality synthetic data by iteratively refining task difficulty using multiple LLMs, with output used for GRPO training.

This paper presents a synthetic data generation method for fine-tuning small LLMs to convert natural language to Cypher queries for property graphs, achieving competitive performance with large proprietary models while enabling local deployment and data sovereignty.

ReGeN is a reference-guided generative pipeline for multivariate time series data that decomposes observed sequences into periodic backbone, stochastic residuals, and cross-variable dependencies to synthesize controllable synthetic data. It demonstrates that generated data can substitute for real data in forecasting tasks, outperforming prior synthetic data generators.