The 4th Financial Narrative Processing Workshop (FNP 2022)

To be held at 13th Edition of the Language Resources and Evaluation Conference (LREC 2022), Marseille, France on 24 June 2022 (full day event).

 

NEW:  The proceedings are now published by LREC: http://www.lrec-conf.org/proceedings/lrec2022/workshops/FNP/index.html

 

 

Workshop Programme: https://www.lancaster.ac.uk/staff/elhaj/FNP2022-programme.pdf 

Workshop Location: Estaque at Palais du Pharo (event took place on 24 June 2022).

 

Keynote Speaker:

Dr Bayan Abu Shawar.

Speaker’s bio, talk title and abstract  👉 http://wp.lancs.ac.uk/cfie/keynote-fnp2022/

 

The talk is available online: Chatbots: are they really useful for financial narratives?

Important Dates:

  • 1st Call for papers & shared task participants: 10 January 2022
  • 2nd Call for papers & shared task participants: 14 February 2022
  • Training set release: 25 February 2022
  • Blind test set release: 25 March 2022
  • Systems submission: 1 April 2022
  • Release of results: 5 April 2022
  • Paper submission deadline: 25 April 2022 (extended)
  • Papers notification of acceptance: 3 May 2022
  • 👉Camera Ready (paper final version): 10 May 2022
  • 👉Registration: Early bird 6 May 2022 (registration URL)
  • Workshop date: 24 June 2022 (full day event)

Workshop Description:

In the domain of Finance, there has been a growing interest in the application of automatic and computer-aided approaches for extracting, summarising, and analysing both qualitative and quantitative financial data. In recent years, manual small-scale research in the Accounting and Finance literature has been scaled up with the aid of NLP, ML, and deep learning methods. For the past 5 years, the FNP workshop has proven to be the leaders in Financial NLP, organising more than six international events as well as introducing NLP and AI shared tasks and providing datasets and methodologies needed to push forward the emerging field of Financial NLP.

Following the success of the Multiling 2019 Financial narrative Summarisation task at RANLP in Bulgaria, the 1st First Narrative Processing workshop (FNP 2018) at LREC’18 in Japan, the 2nd FNP at NoDaLiDa 2019 in Finland, as well as the 1st Joint Workshop on Financial Narrative Processing and MultiLing Financial Summarisation (FNP-FNS 2020) and the 3rd Financial Narrative Processing Workshop (FNP 2021), we have had a great deal of positive feedback and interest in the development of the financial narrative processing field, especially from our shared task participants. FNP 2021 was the largest among our events in terms of the number of attendees. The workshop was run as a hybrid event at Lancaster University in the UK attracting more than 120 registrations with participants coming from 87 cities from 35 countries around the globe. We ran a series of three shared tasks competitions focusing on financial text summarisation (FNS), structure detection (FinTOC), and causality detection (FinCausal) shared tasks. The shared tasks attracted more than 100 teams from different universities and organisations around the globe and resulted in large scale experimental results and state-of-the-art methods and approaches applied to financial textual documents. This shows the importance and growth of this field. 

The FNP 2021 workshop was a 2-day hybrid event with participants physically joining the workshop at Lancaster University and others online through Microsoft Teams. We had 8 renowned invited speakers, and 22 papers presented orally and in a poster session using Gather Town. The shared tasks were hosted on EvalAi where we have an infrastructure to automatically and quickly evaluate the system results. The event was sponsored by Lancaster University’s DSI and Yseop.com. The granted budget was invested to acquire new data through hiring annotators, organise the event (e.g. Gather Town and Eval-Ai leader-board) and distribute money prizes to the three winning teams (each team was awarded €550). Two AI and NLP Financial Firms joined us in organising the workshop (Fortia.fr and Yseop.com) along with 15 organisers from all around the world. Our workshop organising committee is very diverse and gender balanced.

For the 4th edition of FNP at LREC 2022 we will introduce two new languages, Spanish and Greek in addition to the original English and French datasets. The Spanish and Greek datasets will be considered in the summarisation and the document structure shared tasks as below.

 

Sponsor:

FNP 2022 is sponsored by the European Language Resources Association (ELRA). Winners of the shared task will also receive a free registration to attend the FNP 2022 workshop  at LREC 2022 generously provided by ELRA http://www.elra.info/en/.

The free registration can be only used by one of the team members, we’ll get it touch with the winning team and ask for the name of the person attending and presenting the paper.

 

Previous Proceedings

* All FNP proceedings across the years are on ACL Anthology: https://aclanthology.org/venues/fnp/. The 1st FNP was associated with LREC 2018 http://lrec-conf.org/workshops/lrec2018/W27/pdf/book_of_proceedings.pdf 

FNP Google Scholar: https://scholar.google.com/citations?hl=en&user=8Qn7yJ8AAAAJ 

Motivation and topics of interest:

Financial narrative disclosures represent a large part of firms’ overall financial communications with investors. Textual commentaries help to clarify issues obscured by complex accounting methods and footnote disclosures. In addition, narratives summarise corporate strategy, contextualise results, explain governance arrangements, describe corporate social responsibility policy, and provide forward-looking information for investors. They also provide management with an opportunity to obfuscate accounting results and manipulate readers’ perceptions of underlying economic performance. In FNP 2021 we organised a panel of experts to discuss the future of Financial NLP and data leaders from AI firms in France and London. The conclusion was that financial data has increased exponentially in the recent years due to more regulations, this on the other hand has increased the number of financial news surrounding the events of releasing such disclosures. All that together shows the importance of state of the art methodologies that needed to understand and analyse huge and sensitive financial data in a short amount of time.

Topic of Interest in relation to Financial NLP

We encourage work on topics related to analysing financial narratives using state of the art NLP techniques. This includes but is not limited to morphological analysis, disambiguation, tokenization, part-of-speech tagging, named entity recognition, chunking, parsing, semantic role labelling, sentiment analysis, document quality and advanced readability metrics. Across the years the use of NLP and Machine Learning in the financial domain has encouraged studies around gender and ethnicities imbalance, as well as mental health and wellbeing research. Given the focus of the LREC conference, we also encourage research on under-resourced languages and under-represented financial markets. In recent years, FNP has included research on Arabic, Spanish and Portuguese financial markets.

Our collaboration with the MultiLing workshop (http://multiling.iit.demokritos.gr) has highlighted the importance of summarisation across domains and sources that are related to finance (e.g. company blogs, product reviews, market briefs, etc.); this includes financial multilingual and cross-lingual summarisation using single-document summarisation, multi-document summarisation, summarisation evaluation, headline generation, and cross-domain/cross-topic summarisation.

Given the international nature of the event, we particularly welcome FNP papers reporting non-English and multilingual research, describing the different regulatory regimes within which companies operate internationally. Especially that in FNP 2022 we are introducing Spanish and Greek financial datasets in addition to the current English and French datasets.

 

Submit a paper

Shared Tasks

FNP-FNS 2022 shared tasks:

  1. Financial Narrative Summarisation (FNS 2022): summarise financial data in three languages: English, Spanish and Greek. (Register your team here). 
  2. Financial Table of Content Extraction (FinTOC 2022): detect structure of financial documents in three languages: English, French and Spanish (Register your team here). 
  3. Financial Causality Detection (FinCausal 2022): detect causal effects in financial disclosures in English (Register your team here).

For more details about the shared tasks please visit: http://wp.lancs.ac.uk/cfie/shared-tasks/

Organising Committee:

Call For Papers

We invite papers describing original, completed or ongoing, unpublished research. We welcome paper submissions, in theory, methodology, as well as resources and applications in all areas related to Financial Natural Language Processing and Financial Text Analysis. We also welcome submissions on negative results as well as submissions highlighting challenges faced in industrial or academic settings.

We encourage submissions on topics that include, but are not limited to, the following:

  • Applying core technologies on financial narratives: morphological analysis, disambiguation, tokenization, part-of-speech tagging, named entity recognition, chunking, parsing, semantic role labelling, sentiment analysis, document quality and advanced readability metrics, etc.
  • Using NLP to detect misreporting in relation to diversity and wellbeing on issues related to gender, ethnicity, women at work as well as employee mental health and stability.
  • Financial narrative resources: dictionaries, annotated data, tools and technologies etc.
  • Summarisation across domains and sources that are related to finance (e.g. company blogs, product reviews, market briefs, etc.), this includes financial multilingual and cross-lingual summarisation using single-document summarisation, multi-document summarisation, summarisation evaluation, headline generation, cross-domain/cross-topic summarisation.
  • Analysis of Online Social Networks for detection of public opinions towards financial events.
  • Multilingual analysis, describing the different regulatory regimes within which companies operate internationally.
  • Ongoing research and preliminary results
  • Negative results, for example techniques and methodologies that work for certain languages but not on others. Other venues could be showing that state-of-the-art technologies such as BERT could fail on certain tasks or languages.

Share Your LRs!

When submitting a paper from the START page, authors will be asked to provide essential information about resources (in a broad sense, i.e. also technologies, standards, evaluation kits, etc.) that have been used for the work described in the paper or are a new result of your research.
Moreover, ELRA encourages all LREC authors to share the described LRs (data, tools, services, etc.) to enable their reuse and replicability of experiments (including evaluation ones)”.

Paper Submission Instructions:

https://www.softconf.com/lrec2022/FNP/

We follow the same submission instructions as per LREC 2022 guidelines:

Your submission must follow the guidelines, please note that organizers have the right to not include your paper in the final proceedings if it doesn’t follow the guidelines below:

  • consist of 4 to 8 pages, references and appendix excluded (the number of appendix pages should remain reasonable)
  • comply strictly to the LREC stylesheet
  • be formatted as a PDF.

Anti-Harassment Policy:
The workshop supports the ACL anti-harassment policy.

Programme Committee:

  •   Andrew Moore (Lancaster University, UK)
  •   Antonio Moreno Sandoval (UAM, Spain)
  •   Vasiliki Athanasakou (SMU, Canada)
  •   Aris Kosmopoulos (NCSR, Demokritos),
  •   Ahmed AbuRa’ed (Universitat Pompeu Fabra, Spain), 
  •   Nikiforos Pittaras (NCSR, Demokritos).
  •   Denys Proux (Naver Labs, Switzerland)
  •   George Giannakopoulos (SKEL Lab – NCSR Demokritos, Greece)
  •   Houda Bouamor (CMU, Qatar)
  •   Mahmoud El-Haj (SCC, Lancaster University, UK)
  •   Marina Litvak (Sami Shamoon College of Engineering, Israel)
  •   Martin Walker (University of Manchester, UK)
  •   Paul Rayson (SCC, Lancaster University, UK)
  •   Sira Ferradans, France
  •   Dialekti Valsamou (Fortia Financial Solutions, France)
  •   Ismail El Maarouf (Fortia Financial Solution, France)
  •   Steven Young (LUMS, Lancaster University, UK) 
  •   Scott Piao (SCC, Lancaster University, UK)
  •   Mohan Subbiah (LUMS, Lancaster University, UK)
  •   Paulo Alves (Accounting, Universidade Católica Portuguesa, Portugal)
  •   Thomas Schleicher (MBA, Manchester University)
  •   Ana Gisbert (Accounting, UAM, Spain)
  •   John M. Conroy (IDA Center for Computing Sciences, USA)
  •   Elena Lloret (University of Alicante, Spain)
  •   Vangelis Karkaletsis (NCSR Demokritos)
  •   Mark Last (Ben-Gurion University of the Negev, Israel)
  •   Natalia Vanetik (Sami Shamoon College of Engineering, Israel)
  •   George Petasis (NCSR Demokritos, Greece)
  •   Peter Rankel (Elder Research Inc., USA)
  •   Remi Juge (Fortia Financial Solutions, France)
  •   Dominique Mariko (Yseop, France)
  •   Anubhav Gupta (Yseop, France)
  •   Hanna Abi-Akl (Yseop, France)
  •   Hugues de Mazancourt (Yseop, France)
  •   Estelle Labidurie (Yseop, France)
  •   Stephane Durfort(Yseop, France) 

 

List of Accepted papers

Track Paper Number Paper Title Authors
Main workshop 2 FinRAD: Financial Readability Assessment Dataset – 13,000+ Definitions of Financial Terms for Measuring Readability Sohom Ghosh, Shovon Sengupta, Sudip Naskar and Sunny Kumar Singh
FNS Shared Task 3 DiMSum: Distributed and Multilingual Summarization of Financial Narratives Neelesh Kumar Shukla, Amit Vaid, Raghu Chandra Katikeri, Sangeeth Keeriyadath and Msp Raja
FNS Shared Task 5 Transformer-based Models for Long Document Summarisation in Financial Domain Urvashi Khanna, Samira Ghodratnama, Diego Moll´a and Amin Beheshti
Main workshop 6 Discovering Financial Hypernyms by Prompting Masked Language Models Bo Peng, Emmanuele Chersoni, Yu-Yin Hsu, Chu-Ren Huang
Main workshop 8 Sentiment Classification by Incorporating Background Knowledge from Financial Ontologies Timen Stepišnik-Perdih, Andraž Pelicon, Blaž Škrlj, Martin Žnidaršič, Igor Lončarski, Senja Pollak
Main workshop 9 Detecting Causes of Stock Price Rise and Decline by Machine Reading Comprehension with BERT Gakuto Tsutsumi, Takehito Utsuro
Main workshop 10 XLNET-GRU Sentiment Regression Model for Cryptocurrency News in English and Malay Nur Azmina Mohamad Zamani, Jasy Liew Suet Yan, Ahmad Muhyiddin Yusof
FinCausal Shared Task 12 DCU-Lorcan at FinCausal 2022: Span-based Causality Extraction from Financial Documents using Pre-trained Language Models Chenyang Lyu, Tianbo Ji, Quanwei Sun and Liting Zhou
FinTOC Shared Task 15 ISPRAS@FinTOC-2022 Shared Task: Two-stage TOC Generation Model Anastasiia Olegovna Bogatenkova, Oksana Vladimirovna Belyaeva, Andrew Igorevich Perminov and Ilya Sergeevich Kozlov
FNS Shared Task 16 Financial Narrative Summarisation Using a Hybrid TF-IDF and Clustering Summariser: AO-Lancs System at FNS 2022 Mahmoud El-Haj and Andrew Graham Ogden
FinTOC Shared Task 17 GREYC@FinTOC-2022: Handling Document Layout and Structure in Native PDF Bundle of Documents Emmanuel Giguet and Nadine Lucas
FinCausal Shared Task 18 LIPI at FinCausal 2022: Mining Causes and Effects from Financial Texts Sohom Ghosh and Sudip Kumar Naskar
FinTOC Shared Task 19 swapUNIBA@FinTOC2022: Fine-tuning Pre-trained Document Image Analysis Model for Title Detection on the Financial Domain Pierluigi Cassotti, Cataldo Musto, Marco DeGemmis, Georgios Lekkas and Giovanni Semeraro
FinCausal Shared Task 20 iLab at FinCausal 2022: Enhancing Causality Detection with an External Cause-Effect Knowledge Graph Ziwei Xu, Rungsiman Nararatwong, Natthawut Kertkeidkachorn and Ryutaro Ichise
FinCausal Shared Task 21 SPOCK at FinCausal 2022: Causal Information Extraction Using \\Span-Based and Sequence Tagging Models Anik Saha, Jian Ni, Oktie Hassanzadeh, Alex Gittens, Kavitha Srinivas and Bulent Yener
FinCausal Shared Task 22 MNLP at FinCausal2022: Nested NER with a Generative Model Jooyeon Lee, Luan Huy Pham and Özlem Uzuner
FinCausal Shared Task 23 ATL at FinCausal 2022: Transformer based Architecture for Automatic Causal Sentence Detection and Cause-Effect Extraction Abir Naskar, Tirthankar Dasgupta, Sudeshna Jana and Lipika Dey
FinCausal Shared Task 24 ExpertNeurons at FinCausal 2022 Task 2: Causality Extraction for Financial Documents Joydeb Mondal, Nagaraj Bhat, Pramir Sarkar and Shahid Reza
FNS Shared Task 25 Multilingual Financial Documentation Summarization by Team_Tredence for FNS2022 Manish Pant and Ankush Chopra
FNS Shared Task 26 Multilingual Text Summarization on Financial Documents Negar Foroutan, Angelika Romanou, Stéphane Massonnet, Rémi Lebret and Karl Aberer
FNS Shared Task 27 Extractive and Abstractive Summarization Methods for Financial Narrative Summarization in English, Spanish and Greek Alejandro Vaca, Alba Segurado, David Betancur and Álvaro Barbero Jiménez
FNS Shared Task 28 The Financial Narrative Summarisation Shared Task (FNS 2022) Mahmoud El-Haj, Nadhem ZMANDAR, Paul Rayson, Ahmed AbuRa’ed, Marina Litvak, Nikiforos Pittaras, George Giannakopoulos, Aris Kosmopoulos, Blanca Carbajo-Coronado and Antonio Moreno-Sandoval
FinTOC Shared Task 29 The Financial Document Structure Extraction Shared Task (FinTOC 2022) Juyeon Kang, Abderrahim Ait Azzi, Sandra Bellato, Blanca Carbajo Coronado, Mahmoud El-Haj, Ismail El Maarouf, Mei Gan, Ana Gisbert and Antonio Moreno Sandoval