Data Augmentation

4 items

RESEARCHarXiv CS.CL·5/1/2026

Selective Augmentation: Improving Universal Automatic Phonetic Transcription via G2P Bootstrapping

This research proposes Selective Augmentation, a bootstrapping method to improve universal automatic phonetic transcription (APT) by selectively transferring linguistic distinctions to address limited high-quality training data. Exemplified with the MultIPA model, the approach enhanced plosive voicing accuracy by 17.6% and introduced aspiration recognition using data augmented from a helper language like Hindi.

machine learning phonetics Data Augmentation Speech Recognition

RESEARCHarXiv CS.LG·5/1/2026

Fidelity, Diversity, and Privacy: A Multi-Dimensional LLM Evaluation for Clinical Data Augmentation

This research proposes using LLMs (DeepSeek-R1, OpenBioLLM-Llama3, Qwen 3.5) for synthetic mental health data augmentation to address data scarcity and privacy regulations. A comprehensive evaluation framework is introduced, assessing semantic fidelity, lexical diversity, and privacy/plagiarism to mitigate risks like mode collapse or memorization.

synthetic data LLMs security Data Augmentation

RESEARCHarXiv CS.LG·12d ago

IGADA-IoT: IoT Sensor Energy Optimization in Wireless Sensor Networks Driven by Automatic Data Augmentation

This paper proposes IGADA-IoT, an information gap-guided automatic data augmentation framework for IoT sensor energy optimization in wireless sensor networks. It utilizes hierarchical multi-generator collaboration and scheduling to address limitations of existing methods, suchading on single generators.

Sensor Networks IoT Energy Optimization Wireless Sensor Networks

RESEARCHarXiv CS.AI·4/23/2026

Exploring Data Augmentation and Resampling Strategies for Transformer-Based Models to Address Class Imbalance in AI Scoring of Scientific Explanations in NGSS Classroom

This study explores data augmentation strategies to enhance transformer-based models for automated scoring of student scientific explanations, specifically addressing class imbalance. It evaluates methods like GPT-4 generated responses, EASE, and ALP against a SciBERT baseline, using a dataset of 1,466 high school responses.

machine learning natural language processing education technology Data Augmentation