Topic modeling and LLM-assisted analysis of 60k+ juvenile justice opinions identifies 182 topics showing child welfare tripling, punitive declines, vocabulary drift, and risks for AI tools over six decades.
arXiv preprint arXiv:2008.09470 (2020)
13 Pith papers cite this work, alongside 36 external citations. Polarity classification is still indexing.
13
Pith papers citing it
36
external citations · external index
citation-role summary
background 1
method 1
citation-polarity summary
verdicts
UNVERDICTED 13representative citing papers
Large-scale analysis of unfiltered search queries shows geospatial intent at 18% of total, dominated by transactional categories outside traditional GIS scope.
A reproducible framework maps topics from separate corpora into a shared space anchored by the 94 IPTC Media Topics taxonomy via guided BERTopic, centroid scoring, and upward collapse to 17 parent topics.
LLM reasoning refines unsupervised text clusters via coherence checks, redundancy removal, and label grounding, yielding better coherence and human-aligned labels on social media data.
PRISM distills sparse LLM labels into a fine-tuned embedding model for thresholded clustering that separates fine-grained topics better than prior local models or raw frontier embeddings.
Global analysis of LIS papers finds country-specific research method profiles with differences between national and international distributions narrowing over 30 years.
Analysis of 1990-2022 LIS papers via automatic extraction of method entities identifies data resources as the central driver of methodological change exhibiting a cyclical emergence-stability pattern.
BERTopic with contextual augmentation outperforms STM on topic coherence and interpretability for short survey responses, but STM better supports inferential covariate analysis.
TF-IDF identifies labeled experts in the top 25 recommendations 79.5% of the time versus 51.5% for GPT-4o mini on an astronomy observatory dataset.
Bibliometric methods rise from 19.61% to 31.81% usage as LIS scholars age, method diversity increases then declines, and scholars increasingly combine conventional and unconventional methods.
Analysis of three LIS journals finds women authors favor interview methods and men authors favor theoretical approaches across topics, identified via an automatic full-text model.
Granite Embedding Multilingual R2 releases 311M and 97M parameter bi-encoder models that achieve state-of-the-art retrieval performance on multilingual text, code, long-document, and reasoning datasets.
IKMF introduces a dual-stream architecture that converts raw data into semantically rich knowledge via AI mining while maintaining integrity, provenance, and reproducibility through parallel archiving.
citing papers explorer
-
From Punishment to Protection: Charting Six Decades of U.S. Juvenile Justice Through Topic Modeling and LLM-Assisted Analysis
Topic modeling and LLM-assisted analysis of 60k+ juvenile justice opinions identifies 182 topics showing child welfare tripling, punitive declines, vocabulary drift, and risks for AI tools over six decades.
-
Much of Geospatial Web Search Is Beyond Traditional GIS
Large-scale analysis of unfiltered search queries shows geospatial intent at 18% of total, dominated by transactional categories outside traditional GIS scope.
-
A Shared IPTC Topic Space for Cross-Source Topic Modelling
A reproducible framework maps topics from separate corpora into a shared space anchored by the 94 IPTC Media Topics taxonomy via guided BERTopic, centroid scoring, and upward collapse to 17 parent topics.
-
Reasoning-Based Refinement of Unsupervised Text Clusters with LLMs
LLM reasoning refines unsupervised text clusters via coherence checks, redundancy removal, and label grounding, yielding better coherence and human-aligned labels on social media data.
-
PRISM: LLM-Guided Semantic Clustering for High-Precision Topics
PRISM distills sparse LLM labels into a fine-tuned embedding model for thresholded clustering that separates fine-grained topics better than prior local models or raw frontier embeddings.
-
Non-synchronism in Global Usage of Research Methods in Library and Information Science from 1990 to 2019
Global analysis of LIS papers finds country-specific research method profiles with differences between national and international distributions narrowing over 30 years.
-
Data-Driven Evolution of Library and Information Science Research Methods (1990-2022): A Perspective Based on Fine-grained Method Entities
Analysis of 1990-2022 LIS papers via automatic extraction of method entities identifies data resources as the central driver of methodological change exhibiting a cyclical emergence-stability pattern.
-
A Comparative Evaluation of Structural Topic Models and BERTopic for Short, Open-Ended Survey Responses
BERTopic with contextual augmentation outperforms STM on topic coherence and interpretability for short survey responses, but STM better supports inferential covariate analysis.
-
Traditional statistical representations outperform generative AI in identifying expert peer reviewers
TF-IDF identifies labeled experts in the top 25 recommendations 79.5% of the time versus 51.5% for GPT-4o mini on an astronomy observatory dataset.
-
Evolution of Research Method Usage Across the Academic Careers of Library and Information Science Scholars
Bibliometric methods rise from 19.61% to 31.81% usage as LIS scholars age, method diversity increases then declines, and scholars increasingly combine conventional and unconventional methods.
-
Gender Differences in Research Topic and Method Selection in Library and Information Science: Perspectives from Three Top Journals
Analysis of three LIS journals finds women authors favor interview methods and men authors favor theoretical approaches across topics, identified via an automatic full-text model.
-
Granite Embedding Multilingual R2 Models
Granite Embedding Multilingual R2 releases 311M and 97M parameter bi-encoder models that achieve state-of-the-art retrieval performance on multilingual text, code, long-document, and reasoning datasets.
-
Intelligent Knowledge Mining Framework: Bridging AI Analysis and Trustworthy Preservation
IKMF introduces a dual-stream architecture that converts raw data into semantically rich knowledge via AI mining while maintaining integrity, provenance, and reproducibility through parallel archiving.