Candidate papers for the Best Paper Award

1. Conference PC Candidate
   The Conference PC Candidate was selected before the conference by the KICSS 2025 Program Chairs, based on EasyChair review scores, reviewer evaluations, and deliberations, following the traditional selection approach.

The PC nominated the Long Paper titled:

“Beyond Divergent Thinking Assessment: Extending Creative Activity Areas in Creativity Training” by Takahiro Kawaji.

----------------------------------------------------------------

2. Session-Vote Candidate
   
   

Q. How will the Session-Vote winner be determined?

After voting is completed in all 9 technical sessions, we compare the vote share (percentage of votes) for the top paper within each session.
The paper with the highest vote share across all sessions will be selected as the Session-Vote Winner.

For example:

• In Session 1, there are 20 people and they all vote.

  • And paper A receives 12 votes → 60% (12/20)

• In Session 2, 10 people vote.

  • Paper B receives 8 votes → 80% (8/10)

• In Session 3, 25 people vote.

  • Paper C receives 15 votes → 60% (15/25)

Although Paper C in session 3 has the highest number of votes (15), but paper B in session 2 has the highest vote share (80%), so Paper B becomes the Session-Vote Winner.

The following questions were used during the voting process:

Q1: (Verification): Participants were asked to enter the last 4 digits of their KICSS 2025 Registration ID, as printed on the top-left corner of their conference badge, to verify their eligibility..

Q2: Which of the following papers presented in this session (Session ID #: Name ) would you like to nominate as the “Session-Vote Candidate" for Best Paper Award? (Please select only ONE paper)

Session-Vote Candidate Results

Session 1: The Session-Vote Candidate winner from Session 1 was the Long Paper titled: “Simply Incorporating Generative AI into Groups Is Not Enough: Exploring the Effects of Human–AI Collaboration in Groups on Brainstorming” by Koutaro Kamada, Nicharee Manakitrungrueng, and Takaya Yuizono. The paper received 22 votes out of 26 total votes (84.62%) among three presented papers, achieving the highest vote share across all nine sessions, and declared Session-Vote Candidate winner.

Session 2: A total of 22 votes were cast in Session 2. Among the three papers presented, the first presentation, the Long Paper titled: “A Comparative Study of Reward Functions in Learning Prompt-Based Length Control of LLMs via GRPO” by Hayato Fujikoshi and Takeshi Okadome, received 17 out of 20 (77.27%) within Session 2.

Session 3: A total of 11 votes were cast in Session 3. Among the four papers presented, the third presentation, the Long Paper titled: “Manipulating Evaluation Bias in Reasoning LLMs for Idea Evaluation: Approaching Human Criteria” by Hiroaki Furukawa, received 6 out of 11 votes (54.55%).

Session 4: A total of 13 votes were cast in Session 4. Among the four papers presented, the third presentation, the Long Paper titled: “Evaluating the Educational Effectiveness of a Microcontroller Simulator for Beginners” by Shogo Yamagishi, Ren Shimizu, Akitaka Higashi and Megumi Futatsugi, received 7 out of 13 votes (53.85%).

Session 5: A total of 15 votes were cast in Session 5. Among the four papers, the first presentation, the Long Paper titled:  “Conversational AI Agents for Mental Well-Being: A Pilot Randomized Controlled Trial with Afghan Women” by Sofia Sahab, Javaid Haqbeen, Diksha Sapkota and Takayuki Ito, received 8 out of 15 votes (53.33%).

Session 6: A total of 11 votes were cast in Session 6. Among the four papers, the third presentation, the Long Paper titled:  “The third presentation of this session Short Paper titled: “Reconfiguring Collaboration: The Dynamics of "Question-Driven" Scaffolding and Creative Tension in AI-Augmented Sensemaking” by Toshiki Otsuka and Yuki Murata, received 6 out of 11 votes (54.55%).

Session 7: A total of 8 votes were cast in Session 7. Among the four papers, the 2nd presentation, the Long Paper titled:  “HAR-DCWGAN: Dual-Conditional Wasserstein GAN for Human Activity Recognition Data Synthesis” by Shu Rui Jia, Hu Cui and Tessai Hayama, received 6 out of 8 votes (62.50%).

Session 8: A total of 11 votes were cast in Session 8. Among the four papers, the 4th presentation, the Long Paper titled:  “From Output to Process: Benchmarking the Reasoning Patterns of LLMs for AI Risk Scenario Generation” by Arisa Morozumi and Hisashi Hayashi, received 5 out of 8 votes (45.45%).

Session 9: A total of 17 votes were cast in Session 9. Among the three papers, the 1st presentation, the Long Paper titled:  From Argument to Deliberation: Evaluating AI Dialogue with the Discourse Quality Index” by Shun Okuhara and Takayuki Ito, and the 2nd  presentation, the Short Paper Titled: “Financial Inclusion Strengthening in Developing Countries: A Proposed Innovations Policy and System Development in Program Indonesia Pintar” by Sofiana Nurjanah and Sitta Maryam Daniswara, Each paper received 7 out of 17 votes (41.18%), resulting in a tie for the highest vote in Session 9.

----------------------------------------------------------------

3. Conference-Vote Candidate
   During the Conference-Vote Candidate round for the Best Paper Award, 38 attendees were present in the conference hall, and 27 of them cast their votes. All attendees voted for their favorite paper among the 33 papers presented (21 long papers and 12 short papers) across the nine sessions.

The following questions were used during the voting process:

Q1: (Verification): Participants were asked to enter the last 4 digits of their KICSS 2025 Registration ID, as printed on the top-left corner of their conference badge, to verify their eligibility..

Q2: Which paper would you like to nominate as the Conference-Vote Candidate for the Best Paper Award?

Please select one paper from the list of all 33 (21 long papers and 12 short papers) papers presented at KICSS 2025.

Conference-Vote Candidate Results

The Long Paper titled: “Conversational AI Agents for Mental Well-Being: A Pilot Randomized Controlled Trial with Afghan Women”

by Sofia Sahab, Jawad Haqbeen, Diksha Sapkota, and Takayuki Ito received the highest number of votes (4 out of 27) among the 33 listed papers across the nine session and was selected as the Conference-Vote Winner.

This voting process was live-streamed to the conference attendees, and the results were announced immediately in front of them to ensure full transparency.