ATA TEKTalks: Human-Inspired Term Extraction with Sketch Engine

Feeling LOC’d out? ATA TEKTalks is here to help!

This webinar offers a practical introduction to Sketch Engine and explores how it can support freelance translators, in-house linguists, language services providers, and students alike—whether you’re looking to streamline your workflow, sharpen your linguistic insights, or expand your client base.

Sketch Engine is an online text analysis tool that works with large samples of language, called text corpora, to identify what is typical and frequent in a language and what is rare, outdated, going out of use, or what new words or grammar are beginning to be used. In a nutshell, Sketch Engine is a tool to learn how language works.

In current language industry scenarios, Sketch Engine is an ideal method of identifying and extracting clean and reliable data from any source.

Manual extraction is time-consuming. With large quantities of text, automatic term extraction is the only viable option. However, the result can be noisy, contain many non-terms, and require human post-editing.

This webinar will present the principles behind an automatic method that reduces the noise considerably by imitating the more reliable manual term extraction process.

What is ATA TEKTalks?

Language technology, or lang tech for short, can feel like a confusing maze, evolving daily. If you’ve ever felt “LOC’d out” of all the excitement—or just overwhelmed—it’s time to grab your compass and assume your place as a tech-savvy linguist!

ATA’s Language Technology Division is pleased to present ATA TEKTalks, a quarterly series of webinars. ATA TEKTalks provides a platform for leading language technology companies to say hello, offer insights, and answer questions relevant to the work of freelancers, language services providers, and in-house translation departments.

Think of ATA TEKTalks as an open exchange of information where language technology companies and their representatives demonstrate how and why their services can improve your return on investment.

This webinar series is organized in collaboration with ATA’s Language Technology Division.

By the end of this webinar, you will be able to:

1. Conduct corpus queries to explore how words are used across a range of contexts, examining frequency, distribution, and variation.
2. Analyze collocations using Sketch Engine to uncover patterns of co-occurrence that reveal how language works in real, everyday use.
3. Create word sketches that capture a word’s grammatical behavior and typical collocates, offering a concise view of its usage profile.
4. Extract keywords statistically to pinpoint salient terms in a text or corpus, which is useful for domain analysis, terminology work, or content comparison.
5. Support language teaching with authentic, corpus-based examples that enhance textbook materials and make instruction more contextually rich.

About the Presenter

Ondřej Matuška is a linguist and head of sales at Sketch Engine, a corpus system developed by Lexical Computing. His role includes promoting the use of corpora among linguists and language professionals and running training events and workshops. Ondřej also contributes to developing the user interface and improving the user experience of Sketch Engine, the OneClick Terms term extractor and the Lexonomy dictionary writing system.