Workshop Description
For the past seventeen years, the workshops in the TextGraphs series have published and promoted the synergy between the field of Graph Theory (GT) and Natural Language Processing (NLP). The mix between the two started small, with graph-theoretical frameworks providing efficient and elegant solutions for NLP applications. Graph-based solutions initially focused on single-document part-of-speech tagging, word sense disambiguation, and semantic role labeling. They became progressively larger to include ontology learning and information extraction from large text collections. Nowadays, graph-based solutions also target Web-scale applications such as information propagation in social networks, rumor proliferation, e-reputation, multiple entity detection, language dynamics learning, and future events prediction, to name a few.
We plan to encourage the description of novel NLP problems or applications that have emerged in recent years, which can be enhanced with existing and new graph-based methods. We widen the workshop topics beyond the familiar graph domain, encompassing a broader range of less examined structured data domains as well. The seventeenth edition of the TextGraphs workshop aims to extend the focus on exploring rising topics of large language models (LLMs) prompting from the unique perspective of GT. Therefore, our workshop aims to foster stronger, mutually advantageous connections between NLP and structured data, tackling key challenges inherent in each field.
TextGraphs-17 invites submissions on (but not limited to) the following topics:
Knowledge Graphs Meet LLMs. A proper utilization of graph-based methods for reasoning over a Knowledge Graph (KG) is a prospective way to overcome critical limitations of the existing LLMs which lack interpretability and factual knowledge and are prone to the hallucination problem. Vice versa, the incorporation of LLM knowledge learnt from large textual collections may help many graph-related tasks, such as KG completion and graph representation learning. Thus, we are highly interested in novel research on the joint use of KG and LLM for an improved processing of either the NLP or graph domain (preferably both).
Chain Prompting of LLMs. Recent studies show that prompting strategies like Chain-of-Thought and Graph-of-Thought enhance language understanding and generation tasks compared to the traditional few-shot methods. We welcome submissions developing advanced prompting schemes and software for LLMs and other pre-trained machine learning models.
Learning from Structured Data. We greet novel efforts to build a bridge between NLP and various structured data formats including relational and non-relational databases, as well as standardized data formats (such as XML, JSON, RDF, etc.)
Interpretability of NLP Systems. The question of interpretability poses a fundamental challenge for the practical application of NLP methods. We invite researchers to adopt structured data and employ graph-based methods to shed light on decision-making and logic behind modern LLMs. Any work on applying a KG or any other structured knowledge to explore and evaluate factual awareness, treating the interpretability problem from the GT perspective, or any other topic that utilizes graphs and other structured data to make LLMs more understandable, is met with appreciation.
Important Dates
Papers due: May 14, 2024
Notification of acceptance: June 15, 2024
Camera-ready papers due: July 1, 2024
Conference date: August 15–16, 2024
All deadlines are UTC-12; AoE
Submissions
We invite submissions of up to eight (8) pages maximum, plus bibliography for long papers and four (4) pages, plus bibliography, for short papers.
The ACL 2024 templates must be used; these are provided in LaTeX and also Microsoft Word format. Submissions will only be accepted in PDF format.
This year, TextGraph submission is managed through OpenReview. Submit papers by the end of the deadline day (timezone is UTC-12; AoE) via the submission link on our site: https://openreview.net/group?id=aclweb.org/ACL/2024/Workshop/TextGraphs-17
We recommend making authors aware of OpenReview's moderation policy for newly created profiles:
New profiles created without an institutional email will go through a moderation process that can take up to two weeks.
New profiles created with an institutional email will be activated automatically.
Accepted papers will be published by ACL Anthology. See the previous editions here: https://aclanthology.org/venues/textgraphs/
Shared Task
We invite participation in the task of Knowledge Graph Question Answering (KGQA). We will ask the participants to analyze candidate answers with text and graph features. For each query-answer candidate, a graph characterizing paths in Wikidata from entity from the query to the answer entity will be given. See more details at the shared task web page.
Testset and leaderboard are now released! Try to combine LLMs with KGs your way to make the best of text and graph representation by participation!
Organizers
Dmitry Ustalov, JetBrains
Arti Ramesh, Binghamton University
Alexander Panchenko, Artificial Intelligence Research Institute
Yanjun Gao, University of Wisconsin-Madison
Irina Nikishina, University of Hamburg
Andrey Sakhovskiy, Kazan Federal University
Elena Tutubalina, Artificial Intelligence Research Institute
Gerald Penn, University of Toronto
Marco Valentino, Idiap Research Institute
Ricardo Usbeck, University of Hamburg
Contacts
Please direct all questions and inquiries to our official e-mail address (textgraphsOC [at] gmail.com) or contact any of the organizers via their individual emails. You are also invited to join our Telegram group: https://t.me/+kRTCZYTrpJ5jZGVi