Search engines are the workhorses of the World Wide Web, returning billions of responses to billions of queries every day. With the explosion of information on the internet, every website from social media to newsrooms to shopping portals rely on a search engine to help users quickly and easily find information that is of interest, without the need to wade through numerous irrelevant web pages. With the advent of smart assistants like Alexa and Siri, search technology is no longer restricted to a written interface, with more and more users interacting with their devices with voice and gestures. As users interact with the internet in their natural method of communication, it has become important for search engines to understand the different languages of their users. In a global setting, challenges of multilingual data are also faced by backend systems that support search engines like Catalog systems, Ads Servers, Cloud Services, IOT devices and more.
English is a widely used language on the internet. Understandably then, a large proportion of the research in search technologies like language and query understanding systems, has taken place for the English language. These models can be adopted to different languages via transfer learning and domain adaptation. However, it is not scalable to relearn the the model for each new language.
Ensuring that Search works equally well in all languages has several major challenges: How can we properly scale language and query understanding systems to languages that are significantly less wide-spread than English? Can we build universal query understanding models for all languages? How do we serve customers searching in languages with little or no annotated data? How can we leverage state-of-art deep learning research in multilingual language understanding? How can we improve the experience of users searching in a variety of languages? State-of-art NLP research has shown promising progress in building multilingual language understanding models with deep learning and massive amounts of data. We aim to bring together experts from across the globe to share their knowledge and experiences on how to leverage state-of-art science in NLP and deep learning, thus helping achieve an improved search experience in a multilingual setting.
In this first Multilingual Search workshop, we aim to bring together researchers and practitioners from across the world, and, in particular, from different disciplines, such as information retrieval, data mining, machine learning, data science, NLP/NLU, machine translation, transfer learning and other related areas to share their ideas and research achievements in providing a seamless search experience in a multilingual setting.
This workshop will cover the challenges in providing a seamless search experience in a multi-lingual settings. We welcome contributions dealing with all aspects of multilingual search including but not limited to:
Authors are invited to submit papers of 4-8 pages in length. Papers should be submitted electronically in PDF format, using the ACM SIG Proceedings format, with a font size no smaller than 10pt. Submit papers through EasyChair. All submissions will be single blind and peer-reviewed. All accepted papers will be presented at the workshop. In addition, accepted papers will be published in the companion proceedings of the WWW conference and the ACM digital library, unless the authors choose to opt out from publishing their papers. We encourage both academic and industry submissions.
Paper Submission Deadline: March 07, 2021
Acceptance notification: March 26, 2021
Camera-ready due: April 09, 2021
Workshop date: April 15, 2021
Conference dates: April 19-23, 2021
|Session||Talk||Presenter||Duration (min)||PDT (Los Angeles)||EDT (New York)||CEST (Ljubljana)||IST (Mumbai)|
|Introduction||Ashutosh Joshi||5||7:00 am||10:00 am||4:00 pm||7:30 pm|
|Session 1||Invited Talk: Using Translation to Connect People: Low Resource Challenges||Francisco Guzman||30||7:05 am||10:05 am||4:05 pm||7:35 pm|
|Instance Based transfer Learning for Multilingual Deep Retrieval||Andrew Arnold||20||7:35 am||10:35 am||4:35 pm||8:05 pm|
|Query Language Identification with Weak Supervision and Noisy Label Pruning||Sweta Sharma, Vijay Huddar||15||7:55 am||10:55 am||4:55 pm||8:25 pm|
|Break||10||8:10 am||11:10 am||5:10 pm||8:40 pm|
|Session 2||Invited Talk: Multilingual Answer Sentence Ranking via Automatically Translated Data||Alessandro Moschitti||30||8:20 am||11:20 am||5:20 pm||8:50 pm|
|Towards Zero-Shot Learning for Image Retrieval and Tagging||Pranav Agarwal, Ritiz Tambi||20||8:50 am||11:50 am||5:50 pm||9:20 pm|
|Leveraging Multilingual Neural Language Models for On-Device Natural Language Understanding||Huy Tu||15||9:10 am||12:10 am||6:10 pm||9:40 pm|
|Break||10||9:25 am||12:25 am||6:25 pm||9:55 pm|
|Session 3||Invited Talk: Cross Language Speech Retireval||Doug Oard||30||9:35 am||12:35 am||6:35 pm||10:05 pm|
|IndicSOUNDEX Algorithm for text Matching||Christopher DiPersio||15||10:05 am||1:05 pm||7:05 pm||10:35 pm|
|Panel Discussion||Rahul Bhagat, Francisco Guzman, Alessandro Moschitti, Doug Oard||30||10:20 am||1:20 am||7:20 pm||10:50 pm|
|Closing Ramarks||Ashutosh Joshi||5||10:50 am||1:50 am||7:50 pm||11:20 pm|