analiza aiqbench
This dataset provides a massive resource for training speech models on underrepresented domains by leveraging automated pipelines involving large language models for transcription, diarization, and topic labeling. The primary technical contribution lies in the application of AI techniques to expand corpus diversity rather than introducing novel algorithms or quantum methods. Therefore, it falls outside TechPulse's scope as an article where AI is incidental to a broader media study; however, since the core utility described relies entirely on current LLM capabilities for data processing and annotation within computer science (cs.CL), I will classify this under **AI Products** if we consider the dataset itself a product of these tools, or strictly adhere to your "Others" rule because the *subject* is religious broadcasting. Re-evaluating Step 1: Is the PRIMARY subject AI/ML/Quantum? The article describes a corpus of radio broadcasts. The abstract mentions using an automated pipeline and LLMs for processing, but the focus is on the data itself (religious radio transcripts
A large-scale corpus of religious radio broadcast transcripts from webstream recordings in the United States
arXiv:2607.26249v1 Announce Type: new Abstract: Religious radio is a widespread but understudied form of mass communication in the United States, and content-level analysis of it has been constrained by the absence of large-scale transcript data. This Data Descriptor presents a corpus of transcribed English-language religious radio broadcasts captured from live webstreams over a one-month period in July 2025. Fifteen-minute segments were recorded on a rolling schedule from 785 distinct streams, which together rebroadcast the signals of more than two thousand AM and FM stations, yielding over 700,000 recordings and more than 60 million diarized lines of speech. Each recording was transcribed and speaker-diarized with an automated pipeline, and segmented and labeled by programming format and topic using a large language model. The corpus is organized as linked tables of stream metadata, recording metadata, and transcript lines. It supports descriptive study of religious broadcasting across regions and traditions, analysis of how social and political issues are discussed in religious media, and speech-processing research in an underrepresented domain.
Abstract:Religious radio is a widespread but understudied form of mass communication in the United States, and content-level analysis of it has been constrained by the absence of large-scale transcript data. This Data Descriptor presents a corpus of transcribed English-language religious radio broadcasts captured from live webstreams over a one-month period in July 2025. Fifteen-minute segments were recorded on a rolling schedule from 785 distinct streams, which together rebroadcast the signals of more than two thousand AM and FM stations, yielding over 700,000 recordings and more than 60 million diarized lines of speech. Each recording was transcribed and speaker-diarized with an automated pipeline, and segmented and labeled by programming format and topic using a large language model. The corpus is organized as linked tables of stream metadata, recording metadata, and transcript lines. It supports descriptive study of religious broadcasting across regions and traditions, analysis of how social and political issues are discussed in religious media, and speech-processing research in an underrepresented domain.
Computation and Language (cs.CL); Computers and Society (cs.CY); Sound (cs.SD); Applications (stat.AP)
arXiv:2607.26249 [cs.CL]
(or arXiv:2607.26249v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2607.26249 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 264 cuvinte
Subiecte conexe
Bill Gates Will NOT And CANNOT Inject Us With A Microchip In A Vaccine - Fails Basic Fact Check
science20.com · 20 aug. 2026
It’s about ethics in journalism, with Ben Smith
The Verge · 17 aug. 2026
BharatGather: A Culturally-Informed Benchmark Dataset for Misinformation and Fake News Detection in Indian Public Events
arXiv · 4 sept. 2026
Speaker-labeled transcription with WhisperX on SageMaker AI
AWS · 24 sept. 2026