---
title: "SkiLLens — paper by Santo et al. (2026) — summary and ideas"
description: "ESCO, the EU's multilingual taxonomy of skills and occupations, is updated slowly. 27 references."
authors: ["Alessia De Santo", "Lorenzo Malandri", "Fabio Mercorio", "Mario Mezzanzanica", "Navid Nobani"]
year: 2026
venue: "Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)"
source: "https://aclanthology.org/2026.eacl-industry.65/"
doi: "10.18653/v1/2026.eacl-industry.65"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/santo-2026-skillens-recognising-mapping-novel-skills/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/santo-2026-skillens-recognising-mapping-novel-skills/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/17/) / SkiLLens: Recognising and Mapping Novel Skills from Millions of Job Ads Across Europe Using Language Models

# SkiLLens: Recognising and Mapping Novel Skills from Millions of Job Ads Across Europe Using Language Models

**Authors:** Alessia De Santo, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica and Navid Nobani · **Year:** 2026 · **Venue:** [Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)](https://md.research.elamur.ai/free/en/venues/eacl/) ([EACL 2026](https://md.research.elamur.ai/free/en/venues/eacl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.eacl-industry.65/) · [PDF](https://aclanthology.org/2026.eacl-industry.65.pdf) · [DOI 10.18653/v1/2026.eacl-industry.65](https://doi.org/10.18653/v1/2026.eacl-industry.65) · [OpenAlex](https://openalex.org/W7140089569)

**Citations:** 1 citations (OpenAlex, 2026-10-14)

## Summary
ESCO, the EU's multilingual taxonomy of skills and occupations, is updated slowly. As a result, newly demanded competencies such as MLOps or prompt engineering show up in job ads long before they appear in the taxonomy. SkiLLens addresses this with a multilingual human-in-the-loop pipeline built within the PILLARS project. It processes about 18 million recent online job advertisements from the Eurostat Web Intelligence Hub, covering 28 countries and 23 languages.

The pipeline has three stages:
- **Candidate generation:** country-specific word embeddings over national ad corpora are compared with ESCO skills to find possible new skill terms.
- **Filtering:** LLMs and labour-market experts discard irrelevant candidates.
- **Mapping:** confirmed skills are linked to ESCO through embedding-based retrieval, and an LLM picks the best match among the top candidates.

National experts judged more than 70% of the extracted skills to be valid. On skill mapping, the method beats two earlier English-only ESCO-linking approaches. Quality differs noticeably between languages. The work matters because it supports near-real-time, cross-country tracking of emerging skills for education and labour policy.

## Key points
- **Data:** the WIH-OJA NLP sample v3 (69.5M ads, 2018–2023) was used, taking up to 1M of the newest ads per country, about 18M postings in total. Smaller states such as Malta, Cyprus and Estonia contribute far fewer ads.
- **Mapping setup:**
  - ESCO alternative labels were translated into English with Deepseek v3.
  - They were embedded with thenlper/gte-large, the best of the MTEB models tested.
  - The top-3 preferred labels were retrieved from ChromaDB by cosine similarity.
  - GPT-4 then chose the final match.
- **Multilingual evaluation:** 200 alternative labels per language were tested across 22 languages. Acc@3 ranged from roughly 60% (Estonian) to about 90% (Croatian, English).
- **Baselines:** SkiLLens outperforms Decorte et al. (2022) and Clavié & Soulié (2023) on the "tech" and "house" datasets across RP@1/5/10 and MRR.
- **Stated limitations:**
  - Sector coverage of job ads is uneven.
  - Pivot translation into English can lose language-specific nuance.
  - Deciding what counts as a novel skill is inherently ambiguous; inter-annotator agreement across languages is planned future work.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of Milano-Bicocca](https://md.research.elamur.ai/free/en/universities/university-of-milano-bicocca/) — University, Milan, Italy · 15 works · 4,946 citations

📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [💼 Business, Management and Accounting](https://md.research.elamur.ai/free/en/shelves/business-management-and-accounting/) · 32 works · [💼 Organizational Behavior and Human Resource Management](https://md.research.elamur.ai/free/en/shelves/business-management-and-accounting/organizational-behavior-and-human-resource-management/) · 12 works

🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works · [🌍 Multilingual & cross-lingual NLP](https://md.research.elamur.ai/free/en/topics/multilingual-nlp/) · 828 works

## Cite

Cite as: Alessia De Santo, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica and Navid Nobani. 2026. SkiLLens: Recognising and Mapping Novel Skills from Millions of Job Ads Across Europe Using Language Models. Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{santo2026skillens,
  title = {{SkiLLens: Recognising and Mapping Novel Skills from Millions of Job Ads Across Europe Using Language Models}},
  author = {Alessia {De Santo}, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica and Navid Nobani},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 5: Industry Track)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-industry.65},
  url = {https://doi.org/10.18653/v1/2026.eacl-industry.65}
}
```

- [← previous SinFoS: A Parallel Dataset for Translating Sinhala Figures of Speech](https://md.research.elamur.ai/free/en/papers/sofalas-2026-sinfos-parallel-dataset-translating-sinhala/)
- [next → SlugRAG at SemEval-2026 Task 8: Domain-Specific Fine-Tuning and Model Scaling for Multi-Turn RAG Retrieval](https://md.research.elamur.ai/free/en/papers/revankar-2026-slugrag-semeval-2026-task-8/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/17/)

## References

27 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Alessandro Barducci, Simone Iannaccone, Valerio La Gatta, Vincenzo Moscato, Giancarlo Sperlì, and Sergio Zavota. 2022. An end-to-end framework for information extraction from italian resumes. Expert Systems with Applications, 210:118487. [find](https://scholar.google.com/scholar?q=Alessandro+Barducci%2C+Simone+Iannaccone%2C+Valerio+La+Gatta%2C+Vincenzo+Moscato%2C+Giancarlo+Sperl%C3%AC%2C+and+Sergio+Zavota%2E+2022%2E+An+end%2Dto%2Dend+framework+for+information+extraction+from+italian+resumes%2E+Expert+S)
2. Akshay Bhola, Kishaloy Halder, Animesh Prasad, and Min-Yen Kan. 2020. Retrieving skills from job descriptions: A language model based extreme multi-label classification framework. In Proceedings of the 28th international conference on computational linguistics, pages 5832–5842. [find](https://scholar.google.com/scholar?q=Akshay+Bhola%2C+Kishaloy+Halder%2C+Animesh+Prasad%2C+and+Min%2DYen+Kan%2E+2020%2E+Retrieving+skills+from+job+descriptions%3A+A+language+model+based+extreme+multi%2Dlabel+classification+framework%2E+In+Proceedings+of+th)
3. Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2017. [Enriching word vectors with subword information](https://md.research.elamur.ai/free/en/papers/bojanowski-2017-enriching-word-vectors-subword-information/). Transactions of the association for computational linguistics, 5:135–146.
4. Roberto Boselli, Mirko Cesarini, Fabio Mercorio, and Mario Mezzanzanica. 2018. Classifying online job advertisements through machine learning. Future Generation Computer Systems, 86:319–328. [find](https://scholar.google.com/scholar?q=Roberto+Boselli%2C+Mirko+Cesarini%2C+Fabio+Mercorio%2C+and+Mario+Mezzanzanica%2E+2018%2E+Classifying+online+job+advertisements+through+machine+learning%2E+Future+Generation+Computer+Systems%2C+86%3A319%E2%80%93328%2E)
5. Federica Calanca, Luiza Sayfullina, Lara Minkus, Claudia Wagner, and Eric Malmi. 2019. Responsible team players wanted: an analysis of soft skill requirements in job advertisements. EPJ Data Science, 8(1):1–20. [find](https://scholar.google.com/scholar?q=Federica+Calanca%2C+Luiza+Sayfullina%2C+Lara+Minkus%2C+Claudia+Wagner%2C+and+Eric+Malmi%2E+2019%2E+Responsible+team+players+wanted%3A+an+analysis+of+soft+skill+requirements+in+job+advertisements%2E+EPJ+Data+Science%2C+)
6. Benjamin Clavié and Guillaume Soulié. 2023. Large language models as batteries-included zero-shot esco skills matchers. [find](https://scholar.google.com/scholar?q=Benjamin+Clavi%C3%A9+and+Guillaume+Souli%C3%A9%2E+2023%2E+Large+language+models+as+batteries%2Dincluded+zero%2Dshot+esco+skills+matchers%2E)
7. Andrea De Mauro, Marco Greco, Michele Grimaldi, and Paavo Ritala. 2018. Human resources for big data professions: A systematic classification of job roles and required skill sets. Information Processing & Management, 54(5):807–817. [find](https://scholar.google.com/scholar?q=Andrea+De+Mauro%2C+Marco+Greco%2C+Michele+Grimaldi%2C+and+Paavo+Ritala%2E+2018%2E+Human+resources+for+big+data+professions%3A+A+systematic+classification+of+job+roles+and+required+skill+sets%2E+Information+Processi)
8. Jens-Joris Decorte, Jeroen Van Hautte, Johannes Deleu, Chris Develder, and Thomas Demeester. 2022. Design of negative sampling strategies for distantly supervised skill extraction. In RecSys in HR, volume 3218. CEUR. [find](https://scholar.google.com/scholar?q=Jens%2DJoris+Decorte%2C+Jeroen+Van+Hautte%2C+Johannes+Deleu%2C+Chris+Develder%2C+and+Thomas+Demeester%2E+2022%2E+Design+of+negative+sampling+strategies+for+distantly+supervised+skill+extraction%2E+In+RecSys+in+HR%2C+vo)
9. Chuyu Fang, Chuan Qin, Qi Zhang, Kaichun Yao, Jingshuai Zhang, Hengshu Zhu, Fuzhen Zhuang, and Hui Xiong. 2023. Recruitpro: A pretrained language model with skill-aware prompt learning for intelligent recruitment. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Minin… [find](https://scholar.google.com/scholar?q=Chuyu+Fang%2C+Chuan+Qin%2C+Qi+Zhang%2C+Kaichun+Yao%2C+Jingshuai+Zhang%2C+Hengshu+Zhu%2C+Fuzhen+Zhuang%2C+and+Hui+Xiong%2E+2023%2E+Recruitpro%3A+A+pretrained+language+model+with+skill%2Daware+prompt+learning+for+intelligent)
10. Silvia Fareri, Nicola Melluso, Filippo Chiarello, and Gualtiero Fantoni. 2021. Skillner: Mining and mapping soft skills from any text. Expert Systems with Applications, 184:115544. [find](https://scholar.google.com/scholar?q=Silvia+Fareri%2C+Nicola+Melluso%2C+Filippo+Chiarello%2C+and+Gualtiero+Fantoni%2E+2021%2E+Skillner%3A+Mining+and+mapping+soft+skills+from+any+text%2E+Expert+Systems+with+Applications%2C+184%3A115544%2E)
11. Anna Giabelli, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica, and Navid Nobani. 2022. Embeddings evaluation using a novel measure of semantic similarity. Cognitive Computation, 14(2):749– 763. [find](https://scholar.google.com/scholar?q=Anna+Giabelli%2C+Lorenzo+Malandri%2C+Fabio+Mercorio%2C+Mario+Mezzanzanica%2C+and+Navid+Nobani%2E+2022%2E+Embeddings+evaluation+using+a+novel+measure+of+semantic+similarity%2E+Cognitive+Computation%2C+14%282%29%3A749%E2%80%93+763%2E)
12. Anna Giabelli, Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica, and Andrea Seveso. 2020. Neo: A tool for taxonomy enrichment with new emerging occupations. In International Semantic Web Conference, pages 568–584. Springer. [find](https://scholar.google.com/scholar?q=Anna+Giabelli%2C+Lorenzo+Malandri%2C+Fabio+Mercorio%2C+Mario+Mezzanzanica%2C+and+Andrea+Seveso%2E+2020%2E+Neo%3A+A+tool+for+taxonomy+enrichment+with+new+emerging+occupations%2E+In+International+Semantic+Web+Conferenc)
13. Nidhi Goyal, Jushaan Kalra, Charu Sharma, Raghava Mutharaju, Niharika Sachdeva, and Ponnurangam Kumaraguru. 2023. Jobxmlc: Extreme multi-label classification of job skills with graph neural networks. In Findings of the Association for Computational Linguistics: EACL 2023, pages 2181–2191. [find](https://scholar.google.com/scholar?q=Nidhi+Goyal%2C+Jushaan+Kalra%2C+Charu+Sharma%2C+Raghava+Mutharaju%2C+Niharika+Sachdeva%2C+and+Ponnurangam+Kumaraguru%2E+2023%2E+Jobxmlc%3A+Extreme+multi%2Dlabel+classification+of+job+skills+with+graph+neural+networks%2E+)
14. Akshay Gugnani, Vinay Kumar Reddy Kasireddy, and Karthikeyan Ponnalagu. 2018. Generating unified candidate skill graph for career path recommendation. In 2018 IEEE International Conference on Data Mining Workshops (ICDMW), pages 328–333. IEEE. [find](https://scholar.google.com/scholar?q=Akshay+Gugnani%2C+Vinay+Kumar+Reddy+Kasireddy%2C+and+Karthikeyan+Ponnalagu%2E+2018%2E+Generating+unified+candidate+skill+graph+for+career+path+recommendation%2E+In+2018+IEEE+International+Conference+on+Data+Min)
15. Fatih Gurcan and Nergiz Ercil Cagiltay. 2019. Big data software engineering: Analysis of knowledge domains and skill sets using lda-based topic modeling. IEEE access, 7:82541–82552. [find](https://scholar.google.com/scholar?q=Fatih+Gurcan+and+Nergiz+Ercil+Cagiltay%2E+2019%2E+Big+data+software+engineering%3A+Analysis+of+knowledge+domains+and+skill+sets+using+lda%2Dbased+topic+modeling%2E+IEEE+access%2C+7%3A82541%E2%80%9382552%2E)
16. Faizan Javed, Phuong Hoang, Thomas Mahoney, and Matt McNair. 2017. Large-scale occupational skills normalization for online recruitment. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 31, pages 4627–4634. of the European Chapter of the Association for Computational Linguist… [find](https://scholar.google.com/scholar?q=Faizan+Javed%2C+Phuong+Hoang%2C+Thomas+Mahoney%2C+and+Matt+McNair%2E+2017%2E+Large%2Dscale+occupational+skills+normalization+for+online+recruitment%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligenc)
17. V. Napierala, J.; Kvetan and J. Branka. 2022. Assessing the representativeness of online job advertisements. Publications Office of the European Union. [find](https://scholar.google.com/scholar?q=V%2E+Napierala%2C+J%2E%3B+Kvetan+and+J%2E+Branka%2E+2022%2E+Assessing+the+representativeness+of+online+job+advertisements%2E+Publications+Office+of+the+European+Union%2E)
18. A. Tsvetkova and 1 others. 2024. How well do online job postings match national sources in large english speaking countries?: Benchmarking lightcast data against statistical sources across regions, sectors and occupations. OECD Local Economic and Employment Development (LEED) Papers 2024/01, OECD Pu… [find](https://scholar.google.com/scholar?q=A%2E+Tsvetkova+and+1+others%2E+2024%2E+How+well+do+online+job+postings+match+national+sources+in+large+english+speaking+countries%3F%3A+Benchmarking+lightcast+data+against+statistical+sources+across+regions%2C+se)
19. Mike Zhang, Kristian Nørgaard Jensen, and Barbara Plank. 2022a. Kompetencer: Fine-grained skill classification in danish job postings via distant supervision and transfer learning. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 436–447. [find](https://scholar.google.com/scholar?q=Mike+Zhang%2C+Kristian+N%C3%B8rgaard+Jensen%2C+and+Barbara+Plank%2E+2022a%2E+Kompetencer%3A+Fine%2Dgrained+skill+classification+in+danish+job+postings+via+distant+supervision+and+transfer+learning%2E+In+Proceedings+of+t)
20. Mike Zhang, Kristian Nørgaard Jensen, Sif Dam Sonniks, and Barbara Plank. 2022b. Skillspan: Hard and soft skill extraction from english job postings. arXiv preprint arXiv:2204.12811. [find](https://scholar.google.com/scholar?q=Mike+Zhang%2C+Kristian+N%C3%B8rgaard+Jensen%2C+Sif+Dam+Sonniks%2C+and+Barbara+Plank%2E+2022b%2E+Skillspan%3A+Hard+and+soft+skill+extraction+from+english+job+postings%2E+arXiv+preprint+arXiv%3A2204%2E12811%2E)
21. Mike Zhang, Rob van der Goot, and Barbara Plank. 2023. Escoxlm-r: Multilingual taxonomy-driven pre-training for the job market domain. In The 61st Annual Meeting of the Association for Computational Linguistics, pages 11871–11890. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Mike+Zhang%2C+Rob+van+der+Goot%2C+and+Barbara+Plank%2E+2023%2E+Escoxlm%2Dr%3A+Multilingual+taxonomy%2Ddriven+pre%2Dtraining+for+the+job+market+domain%2E+In+The+61st+Annual+Meeting+of+the+Association+for+Computational+L)
22. Kameni Florentin Flambeau Jiechieu and Norbert Tsopze. 2021. Skills prediction based on multi-label resume classification using cnn with model predictions explanation. Neural Computing and Applications, 33(10):5069–5087. [find](https://scholar.google.com/scholar?q=Kameni+Florentin+Flambeau+Jiechieu+and+Norbert+Tsopze%2E+2021%2E+Skills+prediction+based+on+multi%2Dlabel+resume+classification+using+cnn+with+model+predictions+explanation%2E+Neural+Computing+and+Application)
23. Nan Li, Bo Kang, and Tijl De Bie. 2023. Skillgpt: a restful api service for skill extraction and standardization using a large language model. arXiv preprint arXiv:2304.11060. — [arXiv:2304.11060](https://arxiv.org/abs/2304.11060) [find](https://scholar.google.com/scholar?q=Nan+Li%2C+Bo+Kang%2C+and+Tijl+De+Bie%2E+2023%2E+Skillgpt%3A+a+restful+api+service+for+skill+extraction+and+standardization+using+a+large+language+model%2E+arXiv+preprint+arXiv%3A2304%2E11060%2E+%E2%80%94+%5BarXiv%3A2304%2E11060%5D%28htt)
24. Shan Li, Baoxu Shi, Jaewon Yang, Ji Yan, Shuai Wang, Fei Chen, and Qi He. 2020. Deep job understanding at linkedin. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 2145–2148. [find](https://scholar.google.com/scholar?q=Shan+Li%2C+Baoxu+Shi%2C+Jaewon+Yang%2C+Ji+Yan%2C+Shuai+Wang%2C+Fei+Chen%2C+and+Qi+He%2E+2020%2E+Deep+job+understanding+at+linkedin%2E+In+Proceedings+of+the+43rd+International+ACM+SIGIR+Conference+on+Research+and+Develo)
25. Lorenzo Malandri, Fabio Mercorio, Mario Mezzanzanica, and Navid Nobani. 2020. Meet: A method for embeddings evaluation for taxonomic data. In 2020 International Conference on Data Mining Workshops (ICDMW), pages 31–38. IEEE. [find](https://scholar.google.com/scholar?q=Lorenzo+Malandri%2C+Fabio+Mercorio%2C+Mario+Mezzanzanica%2C+and+Navid+Nobani%2E+2020%2E+Meet%3A+A+method+for+embeddings+evaluation+for+taxonomic+data%2E+In+2020+International+Conference+on+Data+Mining+Workshops+%28IC)
26. Raymond J Mooney and Razvan Bunescu. 2005. Mining knowledge from text using information extraction. ACM SIGKDD explorations newsletter, 7(1):3–10. [find](https://scholar.google.com/scholar?q=Raymond+J+Mooney+and+Razvan+Bunescu%2E+2005%2E+Mining+knowledge+from+text+using+information+extraction%2E+ACM+SIGKDD+explorations+newsletter%2C+7%281%29%3A3%E2%80%9310%2E)
27. Niklas Muennighoff, Nouamane Tazi, Loic Magne, and Nils Reimers. 2023. [Mteb: Massive text embedding benchmark](https://md.research.elamur.ai/free/en/papers/muennighoff-2022-mteb-massive-text-embedding-benchmark/). In Proceedings of the 17th Conference

*Updated 31 October 2026*