Upload 2 files

READMEairti.md

@@ -0,0 +1,81 @@
+---
+license: other
+language:
+- en
+tags:
+- wikipedia
+- wikidata
+- arxiv
+- pubmed
+- stackexchange
+- artificial-intelligence
+- multi-source
+- nlp
+size_categories:
+- n<1K
+---
+
+# Artificial Intelligence — Multi-Source Dataset
+
+## Dataset Description
+
+This dataset contains **142 records** about **Artificial Intelligence**, aggregated from **5 different sources**: Wikipedia, Wikidata, arXiv, PubMed, and StackExchange. Each source contributes a different type of content — encyclopedic articles, structured facts, research paper abstracts, medical research abstracts, and community Q&A — giving a broad, multi-perspective view of the topic.
+
+## Source Breakdown
+
+| Source | Records | Content type |
+|---|---|---|
+| Wikidata | 30 | Structured entity facts and descriptions |
+| arXiv | 30 | Research paper titles and abstracts |
+| PubMed | 30 | Medical/health research abstracts |
+| StackExchange | 30 | Community Q&A discussions (Stack Overflow) |
+| Wikipedia | 22 | Full encyclopedic article text |
+
+## Dataset Structure
+
+Each line in the `.jsonl` file is a JSON object with a unified schema across all sources:
+
+| Field | Description |
+|---|---|
+| `source` | Which source this record came from (`wikipedia`, `wikidata`, `arxiv`, `pubmed`, `stackexchange`) |
+| `title` | Title of the article, paper, entity, or question |
+| `url` | Link to the original source page |
+| `summary` | Short summary or description |
+| `content` | Full available text content |
+| `extra` | Source-specific metadata (e.g. Wikidata entity ID, StackExchange score/answer count, PubMed ID) |
+
+### Example records
+
+```json
+{"source": "wikipedia", "title": "Artificial intelligence", "url": "https://en.wikipedia.org/wiki/Artificial_intelligence", "summary": "Artificial intelligence (AI) is the capability of computational systems to perform tasks typically associated with human intelligence...", "content": "...", "extra": {}}
+
+{"source": "stackexchange", "title": "What is the difference between supervised learning and unsupervised learning?", "url": "https://stackoverflow.com/questions/...", "summary": "Score: 293, Answers: 27", "content": "...", "extra": {"score": 293, "answer_count": 27}}
+```
+
+## Source and Licensing
+
+This dataset mixes content from multiple sources, each with its own license. **Attribution and license compliance is the responsibility of the end user for each source used**:
+
+| Source | License |
+|---|---|
+| Wikipedia | CC-BY-SA 4.0 |
+| Wikidata | CC0 (public domain) |
+| arXiv | Varies per paper — check individual paper licenses |
+| PubMed | Varies per article — many are open access, some are not |
+| StackExchange | CC-BY-SA 4.0 |
+
+## Intended Use
+
+This dataset is intended for:
+- Multi-perspective NLP research (comparing encyclopedic, academic, medical, and community framing of the same topic)
+- Fine-tuning or evaluating language models with diverse text styles
+- Building retrieval systems that draw on multiple content types
+- Educational and research use
+
+## Limitations
+
+- **PubMed titles**: some titles include raw citation header text (e.g. journal name and DOI) rather than a clean article title, due to plain-text parsing of PubMed abstracts. Downstream users may want to further clean this field.
+- **No results from Project Gutenberg**: this niche did not return matching free/public-domain books; not all topics will have results from every source.
+- Content quality and depth varies significantly across sources — Wikipedia and Wikidata content is clean and structured, while PubMed and StackExchange content is comparatively raw.
+- Data reflects a single point-in-time snapshot and may not include the latest research or discussions.
+- Licensing is mixed — this dataset should not be treated as a single uniformly-licensed resource.