The Namkin company and Loria - Université de Lorraine invites applications for a postdoctoral position on business event extraction.

Location: Troyes, France and Nancy, France

Application Deadline: 31st January 2024

Starting Date: March 2024

Contract Duration: 1 year (with possible extension)

The industry faces numerous challenges that necessitate the evolution of BtoB marketing tools, in order to develop a valuable offer and provide an enhanced customer experience. Namkin's BrainLab develops industrial marketing tools for digitalizing customer relations, evolving business models, and exploiting business and economic data for business development. One of the key challenges of marketing intelligence is to identify risks and opportunities so as to guide marketing strategies. Among the sources of information useful to detect risks and opportunities, Namkin has identified Business Events, that is, “textually reported real-world occurrences, actions, relations, and situations involving companies and firms” (Jacobs et al., 2018).

The Loria Semagram team specialises in modelling natural language semantics to represent discourse. While modern semantic representations may contain vast quantities of information, they do not always (or necessarily) contain the information that is useful for the concrete application. For instance, significant challenges still persist in dealing with temporal relations and finely-grained negation interpretation.

A number of studies at the crossroads of business intelligence and NLP have focused on the detection or extraction of Business Events (e.g., Arendarenko & Kakkonen, 2012; Han et al., 2018; Jacobs et al., 2018; Jacobs & Hoste, 2020; Jacobs & Hoste, 2022). Despite the richness of the event extraction literature, many challenges still remain. Some of these challenges are concerned with the modelling of the task itself, such as the necessity / benefit of trigger identification for event extraction (see Zhu et al. 2021), some with the scope of the task, such as sentence level vs document level extraction (e.g., Zheng et al. 2019), some with the information necessary to the integration of events in a coherent knowledge base, like factuality detection (e.g., Zhang et al., 2022) and event disambiguation (e.g., Barhom et al., 2019).

Recent research has looked into the benefits of exploiting semantic representations, and in particular Abstract Meaning Representation (AMR; Banarescu et al. 2013), for low-resources scenarios (Huang et al., 2018) and document level event argument extraction (e.g., Xu et al., 2022). However, it appears that AMR has to be adapted in order to optimally support event extraction related tasks (Yang et al., 2023). One major limitation of AMR for document-level event extraction is that AMR works at the sentence level, and thus requires the aggregation of sentence-level representations. AMR is also limited in terms of negation and universal quantification expressive power.

To overcome these issues, we seek to appoint a Postdoctoral Researcher to work on semantic modelling. Some promising new lead was recently provided by Bos (2023) who proposes a new meaning representation system that overcomes expressive power limitations, supports discourse relations and inter-sentential coreferences, and reduces the annotation load. The appointed Postdoctoral Researcher will explore semantic modelling solutions and their application to event extraction in the field of business.

The topic covers various subjects, including: - Computational semantics, - Machine learning with neural networks, - Cross-domain model transfer, - Learning from small data, - Combining top-down (expert-driven) and bottom-up (dataset-driven) models, - Design of meaning representations - Shallow and deep semantic processing and reasoning - Hybrid symbolic and statistical approaches to semantics - Neural semantic parsing - Semantics and ontologies

The successful candidate will be part of Namkin's Data & IA team and the Sémagramme Team at Loria, with co-supervision provided by Agata Marcante and Professor Maxime Amblard.

As part of the role, you will have the opportunity to... - Design, develop and test semantic representation algorithms for text-mining with the aim of identifying significant information in unstructured text. - Collaborate with Namkin’s experts to evaluate the algorithms on real-world use cases. You will be responsible for writing academic papers, technical reports and project deliverables. You will also attend academic conferences or project meetings to present your findings and act as a representative for the team.

Requirements include expertise in semantic representation algorithms, excellent technical writing skills and the ability to work well in a team. * Applicants must hold a PhD in Computer Science, related to Data Systems, Natural Language Processing, or Artificial Intelligence. * They should have proven fluency in at least one programming language, such as Python, R, Java or C++. * Candidates must possess a curious and passionate attitude towards research and learning in general. * Proficiency in French language would be considered a bonus. * Previous experience in the NLP field would be considered advantageous.

How to apply:

send an email to: applications@namkin.fr mailto:applications@namkin.fr

- with the subject starting with ''Namkin-Loria Postdoc'' - with a single PDF attached containing: * Cover letter detailing motivation and qualifications for this position. * Curriculum vitae, with a list of publications and contact details for references.

Interested parties are encouraged to contact us for further information regarding the position before applying.

References

Arendarenko, E., & Kakkonen, T. (2012). Ontology-based information and event extraction for business intelligence. In Artificial Intelligence: Methodology, Systems, and Applications: 15th International Conference, AIMSA 2012, Varna, Bulgaria, September 12-15, 2012. Proceedings 15 (pp. 89-102). Springer Berlin Heidelberg.

Barhom, S., Shwartz, V., Eirew, A., Bugert, M., Reimers, N., & Dagan, I. (2019). Revisiting joint modeling of cross-document entity and event coreference resolution. arXiv preprint arXiv:1906.01753.

Banarescu, L., Bonial, C., Cai, S., Georgescu, M., Griffitt, K., Hermjakob, U., ... & Schneider, N. (2013, August). Abstract meaning representation for sembanking. In Proceedings of the 7th linguistic annotation workshop and interoperability with discourse (pp. 178-186).

Jacobs, G., & Hoste, V. (2020). Extracting fine-grained economic events from business news. In COLING 2020 (pp. 235-245). COLING.

Jacobs, G., & Hoste, V. (2022). SENTiVENT: enabling supervised information extraction of company-specific events in economic and financial news. Language Resources and Evaluation, 56(1), 225-257.

Jacobs, G., Lefever, E., & Hoste, V. (2018). Economic event detection in company-specific news text. In 1st Workshop on Economics and Natural Language Processing (ECONLP) at Meeting of the Association-for-Computational-Linguistics (ACL) (pp. 1-10). Association for Computational Linguistics (ACL).

Han, S., Hao, X., & Huang, H. (2018). An event-extraction approach for business analysis from online Chinese news. Electronic Commerce Research and Applications, 28, 244-260.

Huang, L., Ji, H., Cho, K., Dagan, I., Riedel, S., & Voss, C. (2018, July). Zero-Shot Transfer Learning for Event Extraction. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) (pp. 2160-2170).

Xu, R., Wang, P., Liu, T., Zeng, S., Chang, B., & Sui, Z. (2022). A two-stream AMR-enhanced model for document-level event argument extraction. arXiv preprint arXiv:2205.00241. Yang, Y., Guo, Q., Hu, X., Zhang, Y., Qiu, X., & Zhang, Z. (2023). An AMR-based link prediction approach for document-level event argument extraction. arXiv preprint arXiv:2305.19162.

Zhang, H., Qian, Z., Li, P., & Zhu, X. (2022, November). Evidence-Based Document-Level Event Factuality Identification. In PRICAI 2022: Trends in Artificial Intelligence: 19th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2022, Shanghai, China, November 10–13, 2022, Proceedings, Part II (pp. 240-254). Cham: Springer Nature Switzerland.

Zheng, S., Cao, W., Xu, W., & Bian, J. (2019). Doc2EDAG: An end-to-end document-level framework for Chinese financial event extraction. arXiv preprint arXiv:1904.07535.

Zhu, T., Qu, X., Chen, W., Wang, Z., Huai, B., Yuan, N. J., & Zhang, M. (2021). Efficient document-level event extraction via pseudo-trigger-aware pruned complete graph. arXiv preprint arXiv:2112.06013.

---------------------- Maxime Amblard Université de Lorraine https://members.loria.fr/mamblard https://members.loria.fr/mamblard http://espoir-ul.fr http://espoir-ul.fr/