博士演習・修士実験

Ph.D. or Masters students who take this hands-on exercise program are asked to choose one of the following themes and provide deliverables at the end of the course.

This theme aims to let students accumulate teaching experience on topics relevant to HCI. Students are encouraged to think and design their own tutorials with guidance of the instructor. They are required to give their tutorials to IIS Lab members (mostly, Masters and undergraduates) as a mini-lecture consisting of 2 – 3 classes each of which should be 60 – 90 minutes. Students are strongly recommended to collaborate with each other to deliver a successful tutorial. The topics should be related to intelligent systems with interactivity. Examples (but not limited to) are:

• Machine learning techniques to design intelligent interactive systems
  • Commonly used machine learning methods (e.g., Linear models, Neural networks, SVM, HMM, CRF)
  • Emerging machine learning methods (e.g., deep learning)
• Natural language processing techniques
  • PoS tagging, document classification, summarization, translation, etc.
• Computer vision techniques
  • SLAM, Structure-from-motion, stereo matching, etc.
• Statistical methods
  • Statistical testing methods
  • Use example of statistical tools

All tutorials should focus on how to use algorithms and methods in a realistic interactive system rather than explaining theoretical backgrounds. Thus, tutorials should include many examples showing how to use relevant libraries in languages or platforms IIS Lab members usually use (i.e., iOS, Android, C#, JavaScript, Java, R).

As how I name this theme, the tutorial must be interactive. Do not design an one-way teaching class. :) I mark the design of your tutorial as well as its content.

Given that the course is rather short, we strongly suggest students to choose a focused topic rather than trying to explain broadly. The ultimate goal of this tutorial is to offer attendees practical experience on particular techniques so that they can apply what they have learned to their projects.

All tutorial materials must be in English, but you may deliver them in Japanese. Please coordinate with the instructor to set up date and time for tutorials.

• Your tutorial (1 day, something like 3 hour long)
• Your tutorial materials (slides, code examples, and exercises)

• Week 1: Identifying a possible tutorial topic
• Week 2: Creating a rough structure of your tutorial
• Week 3: Creating the content of your tutorial
• Week 4: Continuing to create the content of your tutorial
• Week 5: Completing the first version of your tutorial
• Week 6: Piloting your tutorial
• Week 7: Revising your tutorial based on your pilot
• Week 8: Executing your tutorial in IIS Lab

This theme aims to prototype applications where human and AI work collaboratively to achieve something for productivity, creativity, or anything that matters to human. You may use any technology though we can offer Open AI API keys.

In particular, you are encouraged to build an extraheric AI system to enhance people's higher-order thinking through human-AI interaction. Please read the following paper carefully before you move forward. https://arxiv.org/abs/2409.09218

That being said, you also have liberty to explore different directions of human-AI interaction. Please actively discuss with Koji to identify the direction you will pursue.

• Your prototype (code)
• Demo video (1-2 mins should suffice.)

• Week 1: Reading the extraheric paper, and identifying a possible application direction
• Week 2: Running a small-scale literature survey
• Week 3: Narrowing down the application domain based on the literature survey
• Week 4: Creating a low-fidelity prototype for the interface
• Week 5: Implementing a prototype
• Week 6: Continuing the implementation
• Week 7: Running a pilot study and revising the implementation
• Week 8: Demonstrating your application

In this theme, students conduct a thorough literature survey on publications on Human-AI Interaction.

You are asked to go through representative papers in Human-AI Interaction as well as examine latest papers published at CHI, UIST, PACM IMWUT, CSCW, and IUI. You are then asked to write a summary of each paper and synthesize future research directions. More specifically,

• Your survey must cover latest work from CHI, UIST, PACM IMWUT, CSCW, and IUI (you may look into multiple conferences if you want).
• You must cover at least 50 papers.

• Your survey paper (10 – 15 page long excluding references in the ACM SIGCHI single-column format).

• Week 1: Identifying a possible survey direction
• Week 2: Running a thorough literature survey
• Week 3: Continuing the literature survey
• Week 4: Performing a thematic analysis on the papers surveyed
• Week 5: Drawing future research opportunities
• Week 6: Writing a survey paper
• Week 7: Continuing writing
• Week 8: Submitting your survey paper