| 日時(Datetime) | 令和7年6月18日(水)4限 (15:10--16:40), 2025/06/18, Wednesday, 4th slot |
|---|---|
| 司会(Chair) | 松井 (Matsui) sensei |
| 講演者(Presenter) | Butaslac Isidro ⅢMendoza, インタラクティブメディア設計学研究室 (Interactive Media Design Research Lab.) |
|---|---|
| 題目(Title) | Augmented Reality in the Everyday: A Human-Centered Approach |
| 概要(Abstract) | As smartphones evolved from luxury gadgets to everyday necessities, they transformed how we communicate, navigate, learn, and work. Today, augmented reality (AR) stands on a similar threshold: full of potential, but still searching for meaningful everyday relevance. In this talk, we explore how AR can become a useful, human-centered tool by focusing on two key functions: training and support. Training systems aim to improve users’ knowledge and skills over time, while support systems provide real-time assistance during tasks. Both play essential roles in shaping AR’s path toward everyday adoption. Just like smartphones became second nature, we want AR to become a natural part of daily life, not just flashy and trendy, but meaningful and useful. |
| 講演者(Presenter) | Wataru Sasaki, ユビキタスコンピューティングシステム研究室 (Ubiquitous Computing Systems Research Lab.) |
|---|---|
| 題目(Title) | Large-Scale Estimation and Analysis of Web Users' Mood from Web Search Query and Mobile Sensor Data |
| 概要(Abstract) | This talk introduces a two-layer framework for nationwide mood sensing that fuses smartphone sensors with web-search behavior. In a 90-day study of 460 volunteers, we captured ten sensor streams, self-reported mood labels, and their Yahoo! Japan queries. A Sensor Mood Model trained on three-hour aggregates reached 72 % accuracy, and its predictions augmented ground-truth data for a Query Mood Model, doubling training size and lifting AUC. Deploying the query model to more than 11 million Yahoo! users yield a three-hourly “Nationwide Mood Score” that climbs on weekends and dips on Mondays. Convergent cross-mapping reveals bidirectional, yet stronger, causal influence from collective mood to COVID-19 case counts, suggesting public sentiment as an early warning signal. The framework demonstrates how personal-scale sensing can bootstrap population-level emotion analytics and opens avenues for real-time epidemiological and social monitoring. Our results underscore the potential of privacy-preserving, low-burden data flows for public-health dashboards. |
| 講演者(Presenter) | Wiraatmaja Christopher, 大規模システム管理研究室 (Large-Scale Systems Management Research Lab.) |
|---|---|
| 題目(Title) | The History of zk-SNARKs and Their Applications Across Multiple Disciplines |
| 概要(Abstract) | The right to privacy is a fundamental human right that has become increasingly important in the digital age. In response, privacy-preserving technologies have emerged as a hot topic across various disciplines in recent decades. Among these technologies, Zero-Knowledge Proofs (ZKPs) stand out as a powerful tool for ensuring data confidentiality and trust. While ZKPs encompass a broad range of techniques, this talk focuses on a particularly practical and scalable variant known as zk-SNARKs. We will explore the historical development of zk-SNARKs, from their initial ideas to real-world implementations, and discuss their transformative potential in fields such as blockchain, artificial intelligence, and beyond. |