{"id":87895,"date":"2026-04-20T15:00:11","date_gmt":"2026-04-20T06:00:11","guid":{"rendered":"https:\/\/www.waseda.jp\/top\/en\/?p=87895"},"modified":"2026-04-21T09:08:25","modified_gmt":"2026-04-21T00:08:25","slug":"an-undergraduate-paper-wins-best-full-research-paper-award-at-a-leading-international-conference-in-information-science","status":"publish","type":"post","link":"https:\/\/www.waseda.jp\/top\/en\/news\/87895","title":{"rendered":"An Undergraduate Paper Wins Best Full Research Paper Award at a Leading International Conference in Information Science"},"content":{"rendered":"

In February 2026, at WSDM 2026 (the ACM International Conference on Web Search and Data Mining) held in the United States, a paper authored by Rikiya Takehi, a fourth-year undergraduate student in the School of Fundamental Science and Engineering*, received the Best Full Research Paper Award.
\n*Academic year at the time of the interview (March 25). Takehi graduated from Waseda University in late March 2026.<\/p>\n

A Novel Perspective on Evaluation Methods for Search Algorithms<\/strong><\/h3>\n

Rikiya Takehi, then a fourth-year student in the Computer Science and Communications Engineering major, School of Fundamental Science and Engineering, was a member of the laboratory led by Professor Tetsuya Sakai (Faculty of Science and Engineering). His research interests include natural language processing, machine learning, reinforcement learning, and information retrieval. In February 2026, Takehi presented his paper entitled \u201cDiversification as Risk Minimization\u201d at the international conference WSDM 2026, where it was selected for the Best Full Research Paper Award.<\/p>\n

WSDM (Web Search and Data Mining) is a premier international conference in the field of information science and is ranked A *\u00a0 in the CORE Rankings*1. With an acceptance rate of approximately 16%, WSDM 2026 selected a single paper for the Best Full Research Paper Award from 799 submissions. It is extremely rare for a paper authored by an undergraduate student to receive this highest distinction. In a post-award interview, Takehi emphasized his deep gratitude toward the mentors who supported him along the way.<\/p>\n

\u201cI am incredibly grateful to Professor Tetsuya Sakai at Waseda University, Professor Fernando Diaz at Carnegie Mellon University, and the researchers at Cornell University and NIST for giving me such valuable opportunities and guidance, despite my limited research experience.\u201d<\/p>\n

\"\"

Rikiya Takehi after receiving the Best Full Research Paper Award<\/p><\/div>\n

Takehi\u2019s award-winning research proposes a new approach to evaluating search algorithms.<\/p>\n

\u201cUsers interact with search engines with a particular intent in mind. For example, when someone searches for \u2018Apple,\u2019 they may be referring either to the fruit or to the global technology company. Since different users have different intents, search engines attempt to present a balanced set of results that reflect multiple possibilities. How search results should be selected and ranked has been a longstanding research question, and my work proposes a new method for addressing it.\u201d<\/p>\n

For queries*2 with multiple possible intents, highly popular intents tend to dominate the top of search rankings, often leaving users with minority intents dissatisfied. This issue stems from ranking results purely by relevance. Diversification algorithms have traditionally been proposed to mitigate this problem. However, Takehi focused on a critical limitation: despite their widespread use, diversification algorithms have not been sufficiently validated from a rigorous mathematical perspective and may fail under certain conditions. His research reexamines these algorithms from a new theoretical viewpoint.<\/p>\n

\u201cUsing simple mathematical transformations and experiments, I showed that diversification algorithms can suffer from issues fundamentally similar to those of relevance-based ranking. To address this gap, my paper proposes a method that directly encodes the true goal of diversification\u2014satisfying users with minority intents\u2014into the objective function itself. If implemented in real-world search engines, this approach could significantly improve user experience.\u201d<\/p>\n

*1: CORE Rankings are maintained by the Computing Research and Education Association of Australasia and provide quality assessments of major computer science conferences. A* is the highest rank
\n*2: Input queries used in information retrieval systems.<\/p>\n

\"\"

Photo taken at the Sakai Laboratory, Nishi-Waseda Campus<\/p><\/div>\n

\u00a0With Plans to Pursue Graduate Studies abroad, He Began Full-Scale Research in His Third year<\/strong><\/h3>\n

Although still an undergraduate, Takehi tackled research challenges that have drawn global attention. Reflecting on his academic path, he explained how his interests developed.<\/p>\n

\u201cMy favorite subjects from a young age were mathematics and physical education. I also spent many years living abroad, so I had a solid foundation in English. After graduating from Waseda University Honjo Senior High School, I wanted to keep the option of working overseas in the future, and I became the first student from my school to enroll in the English-taught degree program of the Department of Computer Science and Communications Engineering (CSCE). I was also considering the School of Political Science and Economics at the time, and I did not enter university with strong technical expertise in computer science.\u201d<\/p>\n

During his early undergraduate years, Takehi devoted much of his time to tennis club activities. He began focusing seriously on research in his third year.<\/p>\n

\u201cIn the winter of my second year, I attended an information session about studying abroad in graduate school, which sparked my interest in U.S. graduate programs. Around the same time, it was time to choose a laboratory for project-based research starting in the third year, and AI was rapidly gaining global attention. I wanted to enter a field where I could maximize my potential, so I decided to pursue AI research and joined Professor Sakai\u2019s lab, given my goal of studying in the U.S.\u201d<\/p>\n

Under Professor Sakai\u2019s supervision, Takehi began research in information retrieval, spending countless days reading academic papers despite having limited prior knowledge. His first paper was accepted at SIGIR-AP*3, and he later received the opportunity to work under researchers at Cornell University, gradually building experience.<\/p>\n

\u201cFrom the fall of my third year, I took a one-year leave of absence to work as a visiting researcher at NIST (the U.S. National Institute of Standards and Technology). At NIST, I conducted research and wrote papers under leading experts in information retrieval. I learned essential skills as a researcher and was often entrusted with entire experiments and algorithm development. This experience greatly expanded my expertise, including exposure to areas such as large language models (LLMs).\u201d<\/p>\n

\"\"<\/p>\n

\"\"

Rikiya Takehi during his stay in the United States<\/p><\/div>\n

After returning to Japan, his work at NIST led to a first-author paper accepted at SIGIR 2025, the top-tier international conference in information retrieval (CORE Rankings A*). Additionally, collaborative research with Cornell University and Hakuhodo Technologies resulted in another first-author paper accepted at ICLR 2025 (CORE Rankings A*). He also participated in a research internship at CyberAgent, the results of which became a jointly authored paper accepted at ICLR 2026, and he pursued collaboration with a professor at Carnegie Mellon University, one of the world\u2019s leading institutions in information science.<\/p>\n

\u201cKnowing that U.S. graduate admissions emphasize research output and letters of recommendation, I focused on turning my ideas into papers presented at international conferences. I was truly honored when Professor Fernando Diaz at Carnegie Mellon evaluated my idea positively. After many twists and turns\u2014including abandoning two research ideas\u2014we started the project in February 2025 and submitted the paper in August. That paper ultimately received the award at WSDM. Insights from my other research experiences also influenced this work.\u201d<\/p>\n

Takehi continues to be highly active, including participating in a research internship at NVIDIA AI, as he prepares for graduate study starting in September.<\/p>\n

\u201cI have received offers from several PhD programs in the United States and am currently deciding where to enroll. Fields such as LLMs are primarily driven by research in the U.S., where collaboration between universities and global tech companies is especially strong. I want to leverage this environment to produce research with worldwide impact. While I envision a career as a researcher in the U.S. after earning my PhD, my long-term goal is to return to Japan and contribute to the advancement of AI technology here.\u201d<\/p>\n

*3: An international conference on information retrieval organized by the Association for Computing Machinery (ACM), primarily focused on the Asia-Pacific region.<\/p>\n

\"\"

At Waseda University, he received the Ono Azusa Memorial Award and a departmental award from the Department of Computer Science and Communications Engineering<\/p><\/div>\n

\u201cWhen in Doubt, Choose the Path That Expands Your Options\u201d<\/strong><\/h3>\n

Finally, we asked Takehi for a message to fellow students at Waseda University.<\/p>\n

\u201cWhat I always kept in mind was this: when in doubt, choose the path that expands your options.\u00a0 I devoted myself fully to the field where I believed I could maximize my potential. Waseda University offers a flexible and supportive research environment that encourages students to take on challenges. By committing wholeheartedly to the field you choose, unexpected opportunities can arise. I hope students will seize those opportunities and take on the world without fear.\u201d<\/p>\n

 <\/p>\n","protected":false},"excerpt":{"rendered":"

In February 2026, at WSDM 2026 (the ACM International Conference on Web Search and Data Mining) held in the Un […]<\/p>\n","protected":false},"author":1,"featured_media":87902,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[81,117],"tags":[45,168,178,302,175],"class_list":["post-87895","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","category-topic","tag-alumni-en","tag-education-en","tag-research-en","tag-student_support-en","tag-students-en"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/posts\/87895","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/comments?post=87895"}],"version-history":[{"count":1,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/posts\/87895\/revisions"}],"predecessor-version":[{"id":87903,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/posts\/87895\/revisions\/87903"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/media\/87902"}],"wp:attachment":[{"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/media?parent=87895"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/categories?post=87895"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.waseda.jp\/top\/en\/wp-json\/wp\/v2\/tags?post=87895"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}