datalyes/retrieval_MIXED

Mixed-Domain Patent Retrieval

Dataset Description

This dataset is part of PatenTEB, a comprehensive benchmark for evaluating text embedding models on patent-specific tasks. PatenTEB comprises 15 tasks across retrieval, classification, paraphrase detection, and clustering, with 2.06 million examples designed to reflect real-world patent analysis workflows.

Paper: PatenTEB: A Comprehensive Benchmark and Model Family for Patent Text Embedding

Task Details

• Task Name: retrieval_MIXED
• Task Type: Retrieval
• Test Samples: 15,574

Symmetric retrieval across partially overlapping technological domains (partial IPC3 overlap). Assesses robustness to partial domain overlaps. Hard negatives from PART_MIX category enable discrimination between related documents.

Dataset Structure

This is a retrieval task where models find relevant patents given a query.

Splits:

• test: Query-document pairs for retrieval evaluation

Columns:

• q
• neg
• pos
• pos_link_type
• neg_computed_domain
• objective
• q_text
• pos_text
• neg_text

Data Sample

Below is a 5-row preview of the test set:

q,neg,pos,pos_link_type,neg_computed_domain,objective,q_text,pos_text,neg_text
129-200-192-504-312,119-472-215-859-05X,030-282-314-526-125,CITES,PART_MIX,MIXED,article slicing method and apparatus [SEP] economically and quickly provide excellent slices such as tomato slices. tomatoes are placed on cutting ...,counter-top type bread cutter with manual type cradle [SEP] slice bread with high efficiency by liftably-fitting a cradle for bread whose transport...,expression method [SEP] a method is provided for obtaining concentrated solutions of mucilaginous material by heating plants in an aqueous alkaline...
157-775-320-337-460,160-018-083-629-805,131-113-556-116-086,CITED_BY,PART_MIX,MIXED,oral hygiene promoting composition containing bamboo salt [SEP] prepare a composition for enhancing oral hygiene excellent in effects on prevention...,fabrication method of oral care composition [SEP] a fabrication method of an oral care composition comprising the steps of drying a cimicifuga root...,helmet [SEP] disclosed is a helmet. in a helmet including a shield having inner coupling parts at both ends thereof fitted and coupled to assembly ...
098-716-479-423-661,116-245-237-950-154,018-128-111-354-589,CITED_BY,PART_MIX,MIXED,"method for manufacturing refined fats and oils [SEP] provide a method for manufacturing refined oil-and-fat with less by-products, good flavor and ...","method for manufacturing refined oil or fat [SEP] provide a production method of refined oil and fat having reduced content of glycidol, 3-chloropr...",rotary screw blancher [SEP] a food processing apparatus includes a pressure vessel defining a compartment having an inlet end for receiving food pr...
100-834-426-928-479,091-042-961-967-461,039-166-649-174-095,CITES,PART_MIX,MIXED,a new method for making krill meal [SEP] prepare a novel krill meal product with superior nutritional and technical properties. this invention comp...,"compositions including krill extracts and conjugated linoleic acid and methods of using same [SEP] methods and compositions for the prevention, the...",waste heat recovery system [SEP] there is provided a waste heat recovery system characterized by a boiler assembly including a heat recovery steam ...
119-627-829-327-043,102-779-009-116-318,041-960-126-976-03X,CITED_BY,PART_MIX,MIXED,"device and method for foaming liquid food, particularly milk [SEP] improve a foam quality of a foamed food product. a device of the present inventi...","method and apparatus for manufacturing milk foam [SEP] provide a method and an apparatus for manufacturing milk foam, the homogeneity and quality o...","magnetic particle scavenging device and method [SEP] the present invention is directed to a device for removing magnetic particles from a liquid, t..."

Evaluation Metrics

This task uses NDCG@10 (Normalized Discounted Cumulative Gain at rank 10) as the primary metric. NDCG measures ranking quality by discounting relevance scores by logarithmic position, normalized by the ideal ranking.

Usage

Load Dataset

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("datalyes/{task_name}")

# Access test split
test_data = dataset['test']

Use with Sentence Transformers

from sentence_transformers import SentenceTransformer

# Load a patent-specialized model
model = SentenceTransformer("datalyes/patembed-base")

# Encode patent texts
embeddings = model.encode(test_data['text'])

Integrate with MTEB

This dataset is designed to be integrated with the MTEB (Massive Text Embedding Benchmark) framework. Integration with MTEB is in progress and will be available once the corresponding pull requests are accepted.

Benchmark Context

This dataset is part of a larger benchmark suite:

Benchmark Component	Description
PatenTEB	15 tasks covering retrieval, classification, paraphrase, clustering
Test Data (Released)	319,320 examples across all 15 tasks
Training/Validation Data	1.74 million examples (planned for future release)
Total Dataset Size	2.06 million annotated instances

Note: Currently, only the test split is publicly available. Training and validation data release is planned for a future date.

All 15 Tasks (NEW to MTEB):

• 3 classification tasks: Bloom timing, NLI directionality, IPC3 classification
• 2 clustering tasks: IPC-based, Inventor-based
• 8 retrieval tasks: 3 symmetric (IN/MIXED/OUT domain) + 5 asymmetric (fragment-to-full)
• 2 paraphrase tasks: Problem and solution paraphrase detection

MTEB Integration: Upcoming (PR in progress)

Citation

If you use this dataset, please cite our paper:

@misc{ayaou2025patentebcomprehensivebenchmarkmodel,
      title={PatenTEB: A Comprehensive Benchmark and Model Family for Patent Text Embedding}, 
      author={Iliass Ayaou and Denis Cavallucci},
      year={2025},
      eprint={2510.22264},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2510.22264}
}

License

This dataset is released under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.

• You are free to share and adapt the material
• You must give appropriate credit
• You may not use the material for commercial purposes
• If you remix, transform, or build upon the material, you must distribute your contributions under the same license

For full license details, see: https://creativecommons.org/licenses/by-nc-sa/4.0/

Contact

• Authors: Iliass Ayaou, Denis Cavallucci
• Institution: ICUBE Laboratory, INSA Strasbourg
• GitHub: github.com/iliass-y/patenteb
• HuggingFace: huggingface.co/datalyes