Scientific literature

1Description

Paper-level records that tie together abstracts, authors and affiliations, venues, citation counts, reference lists, methods sections and figure captions, with a paper-to-patent mapping that shows which patent documents cite or mention each paper. R&D strategy teams, technology scouts, patent analysts and groups training scientific language models use it to follow research fronts, trace how papers turn into inventions and build retrieval and summarization systems.

Subsets

• Abstracts
• Citations
• References
• Methods sections
• Figure captions
• Paper-to-patent mappings

2Schema

3Sample records

Sample record

{
  "title": "Deep Residual Learning for Image Recognition",
  "authors": [
    "Kaiming He",
    "Xiangyu Zhang",
    "Shaoqing Ren",
    "Jian Sun"
  ],
  "venue": "2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",
  "year": 2016,
  "doi": "https://doi.org/10.1109/cvpr.2016.90",
  "arxiv_id": "1512.03385",
  "arxiv_primary_category": "cs.CV",
  "openalex_id": "https://openalex.org/W2194775991",
  "abstract": "Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.",
  "topics": [
    "Advanced Neural Network Applications",
    "Domain Adaptation and Few-Shot Learning",
    "Advanced Image and Video Retrieval Techniques"
  ],
  "cited_by_count": 228971,
  "cited_by_count_as_of": "2026-09-30",
  "reference_count": 81,
  "references_sample": [
    "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (2016)",
    "Understanding the difficulty of training deep feedforward neural networks (2010)",
    "A multigrid tutorial (1987)",
    "Training Very Deep Networks (2015)"
  ],
  "patent_mentions": {
    "method": "Patent documents whose full text, including cited non-patent literature, contains the exact paper title",
    "patent_document_count": 6697,
    "as_of": "2026-10-01",
    "examples": [
      {
        "publication_number": "US20230237649A1",
        "title": "Systems and Methods for Quantification of Liver Fibrosis with MRI and Deep [...]",
        "assignee": "Children's Hospital Medical Center",
        "priority_date": "2020-04-15"
      },
      {
        "publication_number": "US12231559B2",
        "title": "Neural network classifiers for block chain data structures",
        "assignee": "Ledgerdomain Inc.",
        "priority_date": "2019-05-07"
      },
      {
        "publication_number": "US11837354B2",
        "title": "Contrast-agent-free medical diagnostic imaging",
        "assignee": "London Health Sciences Centre Research Inc.",
        "priority_date": "2020-12-30"
      }
    ]
  },
  "source_url": "https://arxiv.org/abs/1512.03385"
}

4Uses

• Technology scouting and research-front mapping
• Linking science to patents for prior-art and licensing work
• Citation-based ranking and expert finding
• Training scientific retrieval, summarization and QA models
• Benchmarking research output of institutions and competitors

6Access

The full Scientific literature dataset (MLC-0501), records on request. Delivered as JSONL, CSV or Parquet under a commercial licence.

Request the full dataset