REPARQA2021: REthinking PAssage Retrieval for Question-Answering November 30, 2021 |
Conference website | https://sites.google.com/view/reparqaworkshopcikm2021/home |
Submission link | https://easychair.org/conferences/?conf=reparqa2021 |
Abstract registration deadline | July 9, 2021 |
Submission deadline | July 15, 2021 |
REPARQA: REthinking PAssage Retrieval for Question-Answering
Research on Question-Answering (QA) systems has recently achieved considerable success in simplified closed-domain settings such as the SQuAD dataset, which provides a preselected passage. Researchers tackled open-domain QA that presents a key challenge in natural language processing (NLP). Open-domain QA considers a large text corpus such as Wikipedia pages instead of a preselected passage for answering a given question. In this context, the Natural Questions (NQ) dataset has presented a more challenging problem. In fact, instead of providing one short passage for each question, NQ gives an entire Wikipedia page which is significantly longer than the passage provided in the other datasets.
An effective open-domain QA system must be able to successfully retrieve the document and the passage on the one hand and comprehend the question context to answer on the other. The current state-of-the-art of deep learning-based research for open-domain QA is often complicated and consist of mainly two components: (1) a passage retriever that selects a small subset of passages from documents (e.g., Wikipedia pages), and then (2) a machine comprehension that examines the retrieved passages to identify the final answer. Several studies showed that the passage retrieval impact could significantly improve answering question tasks.
Several elements are essential for the passage retriever, such as question and passage representation, similarity and attention mechanism between the question and passages, passage ranking techniques, etc.
The REPARQA workshop is the first one that tackles the issue of passage retrieval for open-domain QA. It aims to bring together experts from industry, science, and academia to exchange ideas and discuss ongoing research in open-domain QA and, more precisely, the passage retrieval component. We encourage the description of novel problem definition of passage retrieval for open-domain QA and new datasets in this context. Furthermore, we also encourage contributions developing new techniques for document retrieval for open-domain QA problems.
Traditional research on passage and document retrieval mainly focuses on superficial similarities between the question and the passage (respectively document), such as cosine similarity. The main distinguishing focus of this workshop will be the use of deep neural networks and encoders for passage retrievals, such as the use of encoders to represent questions and passages, integrate attention mechanisms in the passage retrieval framework, etc.
This workshop aims to discover the recent advances in passage retrieval for open-domain QA and improve open-domain QA systems. Thereby, the REPARQ workshop is an opportunity to inspire experts and researchers to share theoretical and practical knowledge of the various aspects of QA systems, to have focused discussions on the topic leading to converting the novel ideas into future innovations.
Submission Guidelines
For this workshop, we encourage submissions that address any aspect of QA-focused passage or document retrieval. We invite submissions of original works that are related -- but are not limited to -- the topics below:
Topics of interests:
- Passage and document representation for open-domain QA
- Attention mechanism for passage retrieval
- Self-attention for passage and document retrieval
- Passage and document retrieval based on unsupervised approaches
- Passage and document ranking for open-domain QA
- Passage and knowledge graph for open-domain QA
- Reinforcement learning for passage retrieval
- Passage and document retrieval using language models
- Graph neural network for passage retrieval
- Graph-based approaches using kernels
- Ensemble approach for passage retrieval
- Semantic understanding of passage and document content
- Entity detection in questions and passages for context retrieval
- Probabilistic graphical models
- Parameterizations of specific passage retrieval systems for open-domain QA
- Impact of passage retrieval performance on overall open-domain QA performance
- Evaluation measures for assessing passage retrieval for open-domain QA
Submission:
We invite two types of submissions, including original research papers (6-8 pages) as well as position papers (2-4 pages). Submissions must be formatted according to the CIKM 2021 conference submissions formatting guidelines. All papers will be peer-reviewed and assessed by the program committee based on their novelty, technical quality, potential impact, clarity, and reproducibility. Participants should publish their source codes and include a link to their repository in the submitted papers. All the papers are required to be submitted via the EasyChair system.
At least one author of each accepted paper must register for the workshop in order to present the paper. For further instructions, please refer to the CIKM 2021 page.
Important dates:
-
July 15, 2021: Paper submission
-
August 15, 2021: Paper acceptance notification
-
Workshop dates will be flexible
Committees
Program Committee
- TBD
Organizing committee
-
Dr. Rafika Boutalbi, University of Stuttgart, Germany
-
Prof. Mohamed Nadif, Université de Paris, CNRS Centre Borelli, FRANCE
-
Rim Hantach, Engie, France
Contact
Dr. Rafika Boutalbi, Analytic computing, IPVS, University of Stuttgart