-
🌱 I’m currently learning the following.
- Mathematical Optimization
- Machine Learning
- Deep Learning (mainly computer vision)
- Quantum Computer (Combinatorial optimization)
- Industrial Application
-
🧑💻 My research topic is about adversarial examples, anomaly detection, and time series forecasting.
- M.S. in Mathematics, Kyushu University(Apr.2020 - Mar.2022 in Japan)
- B.S. in Mathematics, Kyushu University(Apr.2016 - Mar.2020 in Japan)
- Optimization of CLIP Models for Domain-Specific Video Search Kazuya Ueki, Haruki Sato, Yuma Suzuki, Takayuki Hori, Hiroki Takushima, Takumi Takada, Hayato Tanoue, Aiswariya Manoj Kumar, Hiroki Nishihara, Yuki Shibata Proceedings of the 2025 International Conference on Multimedia Retrieval, 2053-2057, 2025
- 植木 一也（明星大），佐藤 開，鈴木 裕真，堀 隆之，宅島 寛貴，髙田 拓実，田之上 隼人，Aiswariya Manoj Kumar，西原 大輝（ソフトバンク）, 大規模映像検索に向けたCLIPモデルの最適化, DIA2025
- Akihiro Yoshida, Haruki Sato, Shiori Uchiumi, Nariaki Tateiwa, Daisuke Kataoka, Akira Tanaka, Nozomi Hata, Yousuke Yatsushiro, Ayano Ide, Hiroki Ishikura, Shingo Egi, Miyu Fuji, Hiroki Kai and Katsuki Fujisawa, Comprehensive and Practical Optimal Delivery Planning System for Replacing Liquefied Petroleum Gas Cylinders, Japan Journal of Industrial and Applied Mathematics, 2024
- Kazuya Ueki, Yuma Suzuki, Haruki Sato, Takayuki Hori, Takumi Takada, Hiroki Takushima, Hayato Tanoue, Aiswariya Manoj Kumar, Hiroki Nishihara, Elevating Video Retrieval Capabilities: A Cross-Modal Approach Utilizing Text and Image Generative Models, CBMI2024
- Takumi Takada, Yuma Suzuki, Hiroki Takushima, Hayato Tanoue, Haruki Sato, Aiswariya Kumar, Hiroki Nishihara, Takayuki Hori, Kazuya Ueki, Direct Metric Optimization for Image Captioning through Reward-Weighted Augmented Data Utilization, ACL2024
- Hayato Tanoue, Hiroki Nishihara, Yuma Suzuki, Takayuki Hori, Hiroki Takushima, Haruki Sato, Takumi Takada, Aiswariya Manoj Kumar, Kazuya Ueki, Softbank-Meisei Submission to the Ego4D Moment Queries Challenge and EgoExo4D Hand Pose Challenge 2024, CVPR2024 workshop
- 植木 一也，鈴木 裕真，佐藤 開，堀 隆之，宅島 寛貴，髙田 拓実，田之上 隼人，Aiswariya Manoj Kumar，西原 大輝，マルチモーダル映像検索 ～テキスト・画像生成モデルを駆使した検索の高精度化～, MIRU2024
- Kazuya Ueki, Yuma Suzuki, Haruki Sato, Takayuki Hori, Hiroki Takushima, Takumi Takada, Hayato Tanoue, Aiswariya Manoj Kumar, Hiroki Nishihara, Integration of Text and Image Generation for Advanced Video Search, ICMLT2024
- Kazuya Ueki, Haruki Sato, Yuma Suzuki, Takumi Takada, Hiroki Takushima, Takayuki Hori, High-Precision Video Retrieval through Image Generation, IEVC2024
- 植木 一也（明星大），佐藤 開，鈴木 裕真，髙田 拓実，宅島 寛貴，堀 隆之（ソフトバンク）, 映像検索の精度向上施策：画像生成モデルを用いたクエリ拡張, https://www.tc-iaip.org/dia/2024/program.html#os1, DIA2024
- Takayuki Hori, Kazuya Ueki, Yuma Suzuki, Hiroki Takushima, Hayato Tanoue, Haruki Sato, Takumi Takada, Aiswariya Manoj Kumar, Waseda_Meisei_SoftBank at Video Browser Showdown 2024, MMM2024
- Keiichiro Yamamura, Haruki Sato, Nariaki Tateiwa, Nozomi Hata, Katsuki Fujisawa, Issa Oe, Hiroki Ishikura, Toru Mitsutake, Diversified Adversarial Attacks based on Conjugate Gradient Method, Thirty-ninth International Conference on Machine Learning(ICML2022), paper link is https://icml.cc/Conferences/2022/Schedule?showEvent=16955.
- Akihiro Yoshida, Haruki Sato, Shiori Uchiumi, Nariaki Tateiwa, Daisuke Kataoka, Akira Tanaka, Nozomi Hata, Yousuke Yatsushiro, Ayano Ide, Hiroki Ishikura, Shingo Egi, Miyu Fujii, Hiroki Kai, Katsuki Fujisawa, Long-Term Optimal Delivery Planning for Replacing the Liquefied Petroleum Gas Cylinder, arxiv preprint, paper link is here.
Task Assignment Problem
The source code can also be found here.
Software Design 2021年12月号
Software Design 2021年11月号
Portofolio Optmization
CUDRA, released: Oct 16, 2021.
A Python library for the job scheduler when using cuda.
デジタル・日本の実現に向けたデータ格付け数理基盤に関するシンポジウム, Dec.2020
The first public seminar "Science Cafe", 17.Sep.2021
NISTが開催した映像解析コンテスト「TRECVID 2023」の映像検索部門と映像説明文生成部門で世界最高水準の精度を達成
ソフトバンクと九州大学が、LPガス容器の配送最適化の共同研究を実施
Data Scientist at SoftBank Corp. April 1, 2022 - Present Tokyo, Japan
Intern at Panasonic Corporation February 12, 2019 - February 22, 2019 Osaka, Japan
- Engaged in software development concerning mathematical optimization using Simulated Annealing (SA).
Amplify Apps Developer at Fixstars Corporation February 2021 - Present Tokyo, Japan
- Developed applications for the Amplify Annealing Engine utilizing the Amplify SDK.
特願2023-157707 (出願中)
特願2021-136392 (公開済み): https://jglobal.jst.go.jp/detail?JGLOBAL_ID=202303009118398320
Nariaki Tateiwa: https://nariaki3551.github.io/index.html
🐱 My GitHub Data
🏆 1,512 Contributions in the Year 2021
📦 72.4 kB Used in GitHub's Storage
🚫 Not Opted to Hire
📜 55 Public Repositories
🔑 6 Private Repositories
I'm a Night 🦉
🌞 Morning 58 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 9.95%
🌆 Daytime 210 commits █████████░░░░░░░░░░░░░░░░ 36.02%
🌃 Evening 275 commits ███████████░░░░░░░░░░░░░░ 47.17%
🌙 Night 40 commits █░░░░░░░░░░░░░░░░░░░░░░░░ 6.86%
📅 I'm Most Productive on Friday
Monday 97 commits ████░░░░░░░░░░░░░░░░░░░░░ 16.64%
Tuesday 86 commits ███░░░░░░░░░░░░░░░░░░░░░░ 14.75%
Wednesday 82 commits ███░░░░░░░░░░░░░░░░░░░░░░ 14.07%
Thursday 82 commits ███░░░░░░░░░░░░░░░░░░░░░░ 14.07%
Friday 144 commits ██████░░░░░░░░░░░░░░░░░░░ 24.7%
Saturday 61 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 10.46%
Sunday 31 commits █░░░░░░░░░░░░░░░░░░░░░░░░ 5.32%
📊 This Week I Spent My Time On
⌚︎ Time Zone: Asia/Tokyo
💬 Programming Languages:
Other 40 hrs 27 mins ██████████████████████░░░ 90.01%
Python 3 hrs 37 mins ██░░░░░░░░░░░░░░░░░░░░░░░ 8.08%
Markdown 22 mins ░░░░░░░░░░░░░░░░░░░░░░░░░ 0.82%
YAML 14 mins ░░░░░░░░░░░░░░░░░░░░░░░░░ 0.53%
CSV 6 mins ░░░░░░░░░░░░░░░░░░░░░░░░░ 0.23%
🔥 Editors:
Browser 38 hrs 7 mins █████████████████████░░░░ 84.82%
VS Code 6 hrs 48 mins ███░░░░░░░░░░░░░░░░░░░░░░ 15.14%
Vim 1 min ░░░░░░░░░░░░░░░░░░░░░░░░░ 0.04%
💻 Operating System:
Mac 39 hrs 7 mins █████████████████████░░░░ 87.06%
Linux 5 hrs 49 mins ███░░░░░░░░░░░░░░░░░░░░░░ 12.94%
I Mostly Code in Python
Python 17 repos ██████████████░░░░░░░░░░░ 56.67%
Jupyter Notebook 5 repos ████░░░░░░░░░░░░░░░░░░░░░ 16.67%
C++ 1 repo ░░░░░░░░░░░░░░░░░░░░░░░░░ 3.33%
HTML 1 repo ░░░░░░░░░░░░░░░░░░░░░░░░░ 3.33%
Rich Text Format 1 repo ░░░░░░░░░░░░░░░░░░░░░░░░░ 3.33%
Timeline
Last Updated on 28/11/2021
