Biography
Know thyself - γνῶθι σεαυτόν
Hi, I’m Yuma Uchiumi, a data science consultant in a consulting company. I was born on December 8, 1997 in Tokyo, Japan. My current interests are the algorithmic marketing and cognitive problems with human thinking.
👻 🍭 🦄 🌈 ⛅ ✨
keywords: Statistical Modeling, Machine Learning, Algorithmic Marketing, Marketing Science, Cognitive Science
Work
Full-time Job
- Data Science Consultant, Accenture (Dec. 2023 – Present)
Accenture Strategy & Consulting Data-AI Group. Japan Tokyo Office.
- Data Science Analyst, Accenture (Apr. 2022 – Nov. 2023)
Accenture Strategy & Consulting Applied Intelligence Group. Japan Tokyo Office.
Part-time Job
- Teaching Assistant, Keio University. (Oct. 2020 – Jan. 2021)
I was TA on a class for the real-time robot control using the physics simulator CopperiaSim in Fall semester 2020.
- Summer Intern in Data Engineering, Yahoo Japan Co., Ltd. (Aug. 2020 – Sep. 2020)
I worked in bigdata analysis at the Finance Office Headquarters Development Division DD Room. I conducted an association analysis on log data on the Web and proposed some marketing measures.
- Data Science Intern, DATAWISE, inc. (Apr. 2020 – Dec. 2020)
I worked on analyzing the GPS data of mobile phone users in data science team. Specifically, I worked on improving the algorithm that estimates the total population from location data of NTT docomo users, crawling and batch processing public data, and building backend apps on GCP.
- R&D Intern, ARAYA, inc. (Jul. 2018 – Mar. 2020)
I was engaged in the R&D project for the computational cost of the Deep Neural Nets. I’ve surveyed related papers and conducted some PoC implementations and experiments on it.
- PwC Analytics Hackathon, PwC Japan Group (Dec. 2020)
Education
- MS in Computer Science (Apr. 2020 - Mar. 2022), Center for Information & Computer Science, School of Science for Open and Environmental Systems, Graduate School of Science and Technology, Keio University, Japan. (Link)
- BE in Computer Science (Apr. 2016 - Mar. 2020), Department of Information & Computer Science, Faculty of Science and Technology, Keio University, Japan. (Link)
Publications
International Conference
- Yuma Uchiumi, Yuki Abe, Michita Imai.
Disentangled Motion Representation Learning for Video Encoding and Generation.
2021 International Symposium on Intelligent Unmanned Systems and Artificial Intelligence (SIUSAI 2021). April 19-21, 2021. Singapore. (Best Student Paper)
Domestic Conference (Japan)
- Yuma Uchiumi, Yosuke Fukuchi, Mitsuhiko Kimoto, Michita Imai.
A Top-down Penetration to the Visual Attention via Elimination of the Ambiguity.
The 35th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI 2021). Jun. 8-11, 2021. Virtual Conference.
- Yuma Uchiumi, Michita Imai.
A Computation of Visual Attentions on the Stochastic Model Using the Fisher Information Matrix.
The 30th Annual Conference of the Japanese Neural Network Society (JNNS 2020). Dec. 2-5, 2020. Virtual Conference.
- Yuma Uchiumi, Yuki Abe, Takuma Seno, Michita Imai.
Unsupervised Motion Feature Extraction in Video via Bi-directional GAN.
The 34th Annual Conference of the Japanese Society for Artificial Intelligence (JSAI 2020). Jun. 9-12, 2020. Virtual Conference.
- Yuma Uchiumi, Yuki Abe, Takuma Seno, Michita Imai.
Motion Disentangled Bidirectional GAN
Human-Agent Interaction Symposium (HAIS 2020). Mar. 6-7, 2020. Virtual Conference.
Theses
- Yuma Uchiumi. (2022). Dynamic Cognitive Models for the Temporal Evolution of Interpretation Results in the Visual Disambiguation. Master’s thesis. Graduate School of Science and Technology, Keio University.
- Yuma Uchiumi. (2020). Disentangled Motion Representation Learning for Video Encoding and Generation. Bachelor’s thesis. Department of Information & Computer Science, Keio University.
Technical Skills
- Programing Languages:
- Python, Ruby, PHP, Java, SQL.
- DevOps Tools:
- Web Frameworks:
- Ruby on Rails, Vue.js, Node.js.
- Cloud Services:
- AWS (EC2, S3, Lambda, Amplify, Sagemaker).
- GCP (ComputeEngine, CloudFunctions, CloudStrage, BigQuerty, AppScript).
Qualifications
- Grade 1, Applied statistics (Natural Science) - Japan Statistical Society Certificate. Japan Statistical Society (JSC). Dec. 2019.
- Big Data Analysis with Scala and Spark - Coursera.org (EPFL). Jun. 2022. 8B2PDQVPPQSP.
- Scaling Python for Data Scientist - DataCamp. May. 2022.
Personality
- Top5 of Clifton Strengths - “Strategic”, “Learner”, “Ideation”, “Individualization”, “Analytical”. (2022)
- GlobeSmart Profile - “Independent”, “Egalitarian”, “Certainty-oriented”, “Direct”, “Task-oriented”. (2022)
- MBTI - INTP-A Type; “Introverted”, “Intuitive”, “Thinking”, “Perceiving”. (2022)
Social Links