Welcome To My Website

Introduction

I am a 1st year masters student in Informatics at the University of Tsukuba. I'm part of the Knowledge and Language Computing Laboratory where I conduct research on Natural Language Processing!

Interests

                      ,                             
                  TZZZ                            
                 1ZZZZI                           
                1ZZZZZl                           
               EZUZZZZW                           
              XZ0,ZZZZZ                           
             OZ3  ZZZZZ             #ZZYEEUZMZZ@  
           .YZL   VZZZZ           7ZZYI    @ZZZO  
           ZZC    UZZZZ          SZZZ      UZZZI  
         .ZZJ     #ZZZZJ        OZZZA      ZZZQ   
        LZZZZZZZZZZZZZZO       ;ZZZY      CZZZm   
       ZZZ         ZZZZT       3ZZZQ      YZZT  ww
      oZU          ZZZZZ       fZZZ      WZZZ% .ZI
     CZO           ZZZZZI       WZZb    3ZZZZ  UX 
 MZZZZZZZZM    LZZZZZZZZZZZ6     mUZZZZZ8 kYZZZI  

I'm currently interested in applying NLP to improve accessibility. The aim of my masters thesis is to develop a computational model that makes Japanese more accessible to non-native speakers. Originally coming from a humanities / social science background, I'm also interested in inter-disciplinary research and aligning AI models to social values.

I host an online kanji dictionary!

Research

Disambiguation of Japanese Heteronyms on Devices with Limited Computational Resources

Non-native speakers of Japanese frequently rely on furigana to read Japanese texts. Furigana is a Japanese reading aid made of small phonetic characters printed above or next to kanji to show how to pronounce them. However, many texts lack furigana, which can make them difficult to read for non-native speakers. Since manually adding furigana to every kanji is tedious, there is a growing need for technology that can automatically add furigana to Japanese text. One major challenge in automatic furigana generation is distinguishing between heteronyms. Heteronyms are words that share the same spelling but have different pronunciations and meanings. For instance, words such as 「辛い」 can have different readings depending on the context, such as "tsurai" (suffer) and "karai" (spicy). This is a significant challenge not only for furigana generation, but also for the development of text-to-speech systems. I am currently conducting research on lightweight programs that can distinguish between different readings of such heteronyms based on the surrounding words.

Through this research, I aim to make Japanese a more accessible and approachable language for a wider range of people.

see more

Projects

picture of a kanji character
Kanji Dictionary

A free online kanji dictionary with a handwriting recognizer

Automatic Furigana

A program which automatically generates furigana for Japanese text

Blog

Recent

20260904 Japan-Michigan Time
20260905 Developer Blog for Kanji Dictionary
See more

Gallery

Taking photos is one of my hobbies. Most photos appearing on this website were taken by me!

See more