52442 research outputs found
Sort by
ExtractGPT: Exploring the potential of Large Language Models for product attribute value extraction
No full textE-commerce platforms require structured product data in the form of attribute-value pairs to offer features such as faceted product search or attribute-based product comparison. However, vendors often provide unstructured product descriptions, necessitating the extraction of attribute-value pairs from these texts. BERT-based extraction methods require large amounts of task-specific training data and struggle with unseen attribute values. This paper explores using large language models (LLMs) as a more training-data efficient and robust alternative. We propose prompt templates for zero-shot and few-shot scenarios, comparing textual and JSON-based target schema representations. Our experiments show that GPT-4 achieves the highest average F1-score of 85% using detailed attribute descriptions and demonstrations. Llama-3-70B performs nearly as well, offering a competitive open-source alternative. GPT-4 surpasses the best PLM baseline by 5% in F1-score. Fine-tuning GPT-3.5 increases the performance to the level of GPT-4 but reduces the model's ability to generalize to unseen attribute values
MBI-KG: A knowledge graph of structured and linked economic research data extrated from the 1937 book "Die Maschinen-Industrie im Deutschen Reich"
Get PDFThe MaschinenBauIndustrie Knowledge Graph (MBI-KG) is a structured and semantically enriched dataset extracted from the 1937 publication “Die Maschinen-Industrie im Deutschen Reich” (The Machinery Industry in the German Reich), published by the “Wirtschaftsgruppe Maschinenbau” and edited by Herbert Patschan. This historical source offers data on German companies within the mechanical engineering industry during the pre-World War II era.
The book was digitized, and Optical Character Recognition (OCR) was applied to extract text. The unstructured extracted data was then structured and semantically enriched to enable data integration and reuse. The semantically enriched data was uploaded into an open-source knowledge-graph software. The resulting knowledge graph includes detailed information about companies, individuals, and administrative entities relevant to the German mechanical engineering industry. The data is accessible through various means, including a SPARQL endpoint, an API, advanced search functionalities, a reconciliation API, and bulk files. Each entity in the knowledge graph can be exported in multiple formats, such as CSV, RDF (ttl), JSON, and NDJSON, ensuring compatibility with diverse research tools and platforms.
This dataset can be reused in various research domains, including economic history, data science, and digital humanities. By providing machine-readable, structured data from a crucial historical period, the MBI-KG facilitates novel analyses and insights into the economic and industrial landscape of early 20th-century Germany. The dataset's interoperability with other data sources and its alignment with FAIR principles further enhance its value for interdisciplinary research and long-term preservatio
How AI helps to compile human intelligence: An empirical study of emerging augmented intelligence for medical image scanning
Get PDF
Zusammenhängende Verträge gem. § 360 BGB und die Reichweite des Widerrufsdurchgriffs unter Berücksichtigung der unionsrechtlichen Vorgaben der Verbraucherrechte-Richtlinie
No full textFairness in algorithmic profiling: The AMAS case
Get PDFWe study a controversial application of algorithmic profiling in the public sector, the Austrian AMAS system. AMAS was supposed to help caseworkers at the Public Employment Service (PES) Austria to allocate support measures to job seekers based on their predicted chance of (re-)integration into the labor market. Shortly after its release, AMAS was criticized for its apparent unequal treatment of job seekers based on gender and citizenship. We systematically investigate the AMAS model using a novel real-world dataset of young job seekers from Vienna, which allows us to provide the first empirical evaluation of the AMAS model with a focus on fairness measures. We further apply bias mitigation strategies to study their effectiveness in our real-world setting. Our findings indicate that the prediction performance of the AMAS model is insufficient for use in practice, as more than 30 of job seekers would be misclassified in our use case. Further, our results confirm that the original model is biased with respect to gender as it tends to (incorrectly) assign women to the group with high chances of re-employment, which is not prioritized in the PES’ allocation of support measures. However, most bias mitigation strategies were able to improve fairness without compromising performance and thus may form an important building block in revising profiling schemes in the present context
