Triple

T15532299
Position Surface form Disambiguated ID Type / Status
Subject Shaoqing Ren E370248 entity
Predicate usedIn P98 FINISHED
Object COCO object detection benchmarks
COCO object detection benchmarks are widely used large-scale evaluation standards for measuring and comparing the performance of object detection algorithms on the COCO dataset.
 E1162637 NE FINISHED

Named-entity recognition

Before disambiguation, gpt-5-mini classified whether the object phrase is a named entity — the step behind the object's NE type shown above.

Instruction
Given a phrase, classify it is english named entity (e.g., persons, organizations, works of art) in Latin script, or not (e.g., literals, dates, URLs, verbose phrases). For disambiguation, the statement where the phrase occurs as object is also given. Please return a JSON object with `phrase` (string, the phrase being analyzed) and `is_ne` (boolean, indicating whether the phrase is a Named Entity).
Input
Phrase: COCO object detection benchmarks | Statement: [Shaoqing Ren, usedIn, COCO object detection benchmarks]

Disambiguation candidates (2 decisions)

The exact options the model was shown at each disambiguation step, with the option it chose highlighted — the evidence behind this triple's disambiguated ids.

NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07 
Target entity: COCO object detection benchmarks
Context triple: [Shaoqing Ren, usedIn, COCO object detection benchmarks]
  • A. R-CNN
    R-CNN is a pioneering deep learning framework for object detection that combines region proposals with convolutional neural networks to accurately localize and classify objects in images.
  • B. KeypointRCNN
    KeypointRCNN is a deep learning model architecture in PyTorch’s torchvision library designed for object detection combined with human pose estimation via keypoint prediction.
  • C. FasterRCNN
    FasterRCNN is a popular two-stage object detection architecture that first proposes candidate regions and then classifies and refines bounding boxes, widely used in computer vision tasks.
  • D. MaskRCNN
    MaskRCNN is a deep learning model architecture for instance segmentation that extends Faster R-CNN by adding a branch to predict segmentation masks for individual objects in an image.
  • E. YOLO
    "YOLO" is a comedic hip-hop song and music video by The Lonely Island, featuring Adam Levine and Kendrick Lamar, that parodies the phrase "you only live once" by humorously promoting extreme caution and risk avoidance.
  • F. None of above. chosen
  • G. Unsure - the case is ambiguous/there is not enough information to decide.
NED2 Entity disambiguation (via description) gpt-5-mini-2025-08-07 
Target entity: COCO object detection benchmarks
Target entity description: COCO object detection benchmarks are widely used large-scale evaluation standards for measuring and comparing the performance of object detection algorithms on the COCO dataset.
  • A. R-CNN
    R-CNN is a pioneering deep learning framework for object detection that combines region proposals with convolutional neural networks to accurately localize and classify objects in images.
  • B. KeypointRCNN
    KeypointRCNN is a deep learning model architecture in PyTorch’s torchvision library designed for object detection combined with human pose estimation via keypoint prediction.
  • C. FasterRCNN
    FasterRCNN is a popular two-stage object detection architecture that first proposes candidate regions and then classifies and refines bounding boxes, widely used in computer vision tasks.
  • D. MaskRCNN
    MaskRCNN is a deep learning model architecture for instance segmentation that extends Faster R-CNN by adding a branch to predict segmentation masks for individual objects in an image.
  • E. YOLO
    "YOLO" is a comedic hip-hop song and music video by The Lonely Island, featuring Adam Levine and Kendrick Lamar, that parodies the phrase "you only live once" by humorously promoting extreme caution and risk avoidance.
  • F. None of above. chosen

How the object was described

The object's one-sentence description was generated by prompting gpt-5.1 with the object name and this triple as context.

Instruction
Generate a one-sentence description of the target entity. 
You are given a context triple in the form (subject, predicate, object), where the object is the target entity. 
# Instructions
Use the triple to infer relevant information about the entity. Describe the entity based on what is most defining, well-known. 
Avoid repeating the information from the triple, unless really essential.
# Response Format
Return only the sentence: "Description: [one-sentence description of the target entity]"
Input
Entity: COCO object detection benchmarks
Triple: [Shaoqing Ren, usedIn, COCO object detection benchmarks]
Generated description
COCO object detection benchmarks are widely used large-scale evaluation standards for measuring and comparing the performance of object detection algorithms on the COCO dataset.

Provenance (5 batches)

Stage Batch ID Job type Status
creating batch_69d85cc521a08190921fb50319dddc34 elicitation completed
NER batch_69e0414877d88190804ee76566004e13 ner completed
NED1 batch_69ff3d5e82a48190bb0a10ebc2412129 ned_source_triple completed
NED2 batch_69ff413a68488190a6c8907e36a602dc ned_description completed
NEDg batch_69ff3ea7d5ac81908bd1ee64de39dba7 nedg completed
Created at: April 10, 2026, 4:06 a.m.