Patent Task for WAT2024

[HOME]

TASK OVERVIEW

The WAT2024 patent task uses JPO Patent Corpus (JPC).

JPO PATENT CORPUS

JPO Patent Corpus (JPC) was constructed by the Japan Patent Office (JPO). This corpus consists of a Chinese-Japanese, Korean-Japanese, and English-Japanese patent description corpus of 1M parallel sentences. Most setences in the corpus were from four International Patent Classification (IPC) sections: Cemistry (Ch), Electricity (El), Mechanical engineering (Me), and Physics (Ph).

From 2022, the ko-ja test-N2 set was removed due to a technical problem, new test-N4 sets have been added, and test-2022 sets have been updated instead of previous test-N sets. (The previous patent tasks at WAT2018--2021 .)

test-2022: The union of the following four sets.

test-N1: Patent documents from patent families published between 2011 and 2013. These are the same data as the test sets used in the past years' WAT.

test-N2: Patent documents from patent families published between 2016 and 2017. (except for ko-ja)

test-N3: Patent documents published between 2016 and 2017. Target sentences are manually created by translating source sentences.

test-N4: Patent documents from patent families published between 2019 and 2020.

Corpus statistics:

Language Pair	Data Type	File Name	Size	Sections:Ratios	Published Years	Sentence Alignment
ZH<-->JA	TRAIN	train.{zh,ja}	1,000,000	Ch/El/Me/Ph:25%/25%/25%/25%	2011-2013	Automatic
	DEV	dev.{zh,ja}	2,000
	DEVTEST	devtest.{zh,ja}	2,000
	TEST	test-n1.{zh,ja}	2,000
	TEST	test-n2.{zh,ja}	3,000	Ch/El/Me/Ph:Unknown	2016-2017	Automatic
	TEST	test-n3.{zh,ja}	204	Ch/El/Me/Ph:Unknown	2016-2017	Manual
	TEST	test-n4.{zh,ja}	5,000	Uncontrolled	2019-2020	Automatic
	TEST	test-2022.{zh,ja}	10,204	Uncontrolled	2011-2020	Automatic/Manual
KO<-->JA	TRAIN	train.{ko,ja}	1,000,000	Ch/El/Me/Ph:25%/25%/25%/25%	2011-2013	Automatic
	DEV	dev.{ko,ja}	2,000
	DEVTEST	devtest.{ko,ja}	2,000
	TEST	test-n1.{ko,ja}	2,000
	TEST	test-n2.{ko,ja}	0	N/A
	TEST	test-n3.{ko,ja}	230	Ch/El/Me/Ph:Unknown	2016-2017	Manual
	TEST	test-n4.{ko,ja}	5,000	Uncontrolled	2019-2020	Automatic
	TEST	test-2022.{ko,ja}	7,230	Uncontrolled	2011-2020	Automatic/Manual
EN<-->JA	TRAIN	train.{en,ja}	1,000,000	Ch/El/Me/Ph:25%/25%/25%/25%	2011-2013	Automatic
	DEV	dev.{en,ja}	2,000
	DEVTEST	devtest.{en,ja}	2,000
	TEST	test-n1.{en,ja}	2,000
	TEST	test-n2.{en,ja}	3,000	Ch/El/Me/Ph:Unknown	2016-2017	Automatic
	TEST	test-n3.{en,ja}	668	Ch/El/Me/Ph:Unknown	2016-2017	Manual
	TEST	test-n4.{en,ja}	5,000	Uncontrolled	2019-2020	Automatic
	TEST	test-2022.{en,ja}	10,668	Uncontrolled	2011-2020	Automatic/Manual

HOW TO OBTAIN JPC

Complete the license agreement (English/Japanese). (Neither stamp nor signature is essential.)
Email the agreement to the Japan Patent Office (PA0630 -at- jpo.go.jp).
WAT organizers/JPO staffs will email to notify the applicant of a link to download this corpus, once the JPO receives the original of the agreement and approves the application.

SUBMISSION SITE

Translation result submission site (JPC3)

CONTACT

For questions, comments, etc. please email to "wat-organizer -at- googlegroups -dot- com".

CHANGE LOG

2023-04-23: Updated for WAT2024.
2023-04-20: "HOW TO OBTAIN" was updated.
2022-05-18: "HOW TO OBTAIN" was updated.
2022-04-27: Training data size shown in corpus statistics was modified.
2021-03-30: Site opened.

NICT (National Institute of Information and Communications Technology)
Kyoto University
Last Modified: 2023-04-20