Research Area
-
Natural Language Processing
We are presently engaged in research and development across a broad spectrum in this field, from fundamental technologies such as morphological analysis, named-entity recognition, and dependency analysis, to advanced applications like text mining, information retrieval, information extraction, and conversational processing. We are also actively utilizing natural language processing in machine learning, large-scale data analysis, and crowdsourcing.
Member
-
Speech Processing
LY Corporation is conducting research and development, aiming to create novel value to our services through advanced voice user interface and acoustic analysis of music and video data. In the area of Voice UI, we've been developing speech recognition and synthesis technologies to enhance user experiences, mainly in the search and navigation functions of our services. In the area of acoustic analysis of music and video data, we are developing a wide range of service functions, such as music recommendation systems and fraud detection in User-Generated Content (UGC) video, to improve operational efficiency.
Member
-
Image Processing
At LINE Yahoo, we handle an enormous volume of image and video data across a variety of services, including advertising, e-commerce, video SNS, and news media. Aiming to enhance service quality and deliver new user experiences, we are dedicated to the research and development of image and video processing technologies.
Member
-
Multimodal
Generative AI has been dominated by single-modal conversions, such as text-to-image and image-to-text. In recent years, research and development of technologies that comprehensively handle multimodal input, including images, text, and voice, has progressed. This evolution of multimodal AI makes it possible to handle information in a multifaceted, intuitive, and high-resolution manner by connecting media and people, and even between media. We aim to utilize this technology to create new services that are even more convenient and useful than conventional services.
Member
-
Information Retrieval
Information retrieval systems that people use every day are like doors that lead to everywhere in the world, and it is hard for people today to imagine life without a search engines. Such information retrieval systems are facing major changes due to advances in AI technology in recent years. In LY Corporation, we are actively engaged in research and development in order to revamp search experiences on our Web, Chiebukuro and shopping search services and to guarantee our users accurate, effective and fair information access. We are training retrieval models of machine learning ranking as well as those based on recent large language models (LLMs). We are also developing a search platform that combines a conventional search of sparse vectors with an approximate kNN search for vector representations of generative language models. As for the applications of the latest LLMs, we are tackling the "Retrieval Augmented Generation" that alleviates "hallucination" problems and promotes the generation of evidence-grounded answers. We will further advance the idea of allowing not only humans but also AI to freely use information retrieval and give all AI, including LLMs, the power to utilize information retrieval at both training and inference stages; thus Retrieval Enhanced AI systems dramatically expands their problem-solving ability, processing power and explainability.
Member
-
Machine Learning & Data Science
Machine learning is a research field aimed at learning general patterns from data, and it is applied in many fields such as natural language processing, information retrieval, voice processing, and image processing. At LY Corporation, we not only develop new machine learning methods but also apply these techniques to enhance real-world services such as advertising and recommendations, and to extract valuable insights from the vast and diverse data generated by our numerous users. Starting with the application of causal inference and counterfactuals to model training based on transaction log data, addressing challenges in real-world scenarios with large-scale traffic and data generation is a unique contribution in this field where industry and academia are closely connected.
Member
-
Generative AI
Generative AI refers to artificial intelligence capable of creating various types of content, such as language and images. Unlike traditional AI, which primarily focuses on classification and search, generative AI has its ability to produce creative content comparable to human output, and thus has attracted significant attention. Research on generative AI in the field of language and images spans a wide range, from fundamental studies to the enhancement of LY Corporation services.
Member
-
Trustworthy AI
Trustworthy AI is a challenging research field that grapples with many unresolved issues, including the safety, fairness, and transparency of AI. We are working on establishing methods to make and validate AI, including large-scale language models, to ensure they are helpful, harmless, and honest. We are also researching how to protect human rights in an AI-powered society.
Member
-
Security & Privacy
We are engaged in research and development of technologies enhancing security and privacy such as authentication, trust management, and privacy-preserving machine learning towards utilizing personal data in a trustworthy way. We have been involved in international standardization activities for authentication technologies with FIDO Alliance to advance identity management that ensures security and usability without the reliance on passwords. We also contribute to LY's services through on our research achievements such as differentially private federated learning for Sticker Auto-suggest in LINE app.
Member
-
Human-Computer Interaction
We are engaged in researching methods to improve the usability of widely used devices such as smartphones and PCs. We also strive to enhance the user-friendliness of advanced technologies like virtual reality and voice recognition. This involves analyzing the user interfaces of existing devices from the perspectives of ergonomics and experimental psychology and developing innovative devices and interaction methods. Human-Computer Interaction is a multidisciplinary research field, and we collaborate closely with colleagues from various departments within our corporation. For instance, we utilize crowdsourcing to gather extensive smartphone usage data and work alongside experts in security to explore the convenience of security technologies. In addition to pursuing advanced interaction technologies, we conduct numerous experiments to assess whether people can effectively use these technologies. Our goal is to provide services that many users can comfortably use in the future.
Member
-
Crowdsourcing
The act of creating necessary services, ideas, and content by soliciting contributions from a large and unspecified group of people is known as crowdsourcing, a term coined by journalist Jeff Howe in 2006 as a portmanteau of "crowd" and "sourcing." With the expansion of online communities in recent years, there has been growing attention to services that utilize the power of the crowd across various fields. Our company launched its crowdsourcing service, Yahoo! Crowdsourcing, in January 2013. This service allows companies to resolve their "tasks" with the help of numerous users, who can earn points as a reward. We quickly acquired over 200,000 registered users, making Yahoo! Crowdsourcing one of the largest services of its kind in Japan in less than a year after its inception. Crowdsourcing is a research area that intersects multiple disciplines, such as education, economics, psychology, and computer science, because it lies at the intersection of the crowd and machines, and work and leisure. We aim to contribute to solving societal challenges by evolving crowdsourcing into a more sophisticated system of division of labor through research that investigates and experiments with new tasks that can be addressed by crowdsourcing.
Member
