Publications

Advancements in virtual reality (VR) technology have enabled us to engage in embodied communication within metaverse environments as part of our daily interactions. This evolution has spurred research into the "Proteus Effect," a phenomenon where individuals' self-perception and perception of others are influenced by their avatars, leading to behavioral changes. In this study, we explored how metaverse platforms can facilitate creative communication in workplace settings by enabling diverse individuals to share opinions more openly. Focusing on gender bias, we examined the impact of using avatars of a different gender during male-female interactions on speech and receptivity. Our findings indicate that gender-swapped avatars can influence communication dynamics, potentially enhancing openness.

本研究では、キーワード共起ネットワークを用いた会議ダイナミクスの評価に関する軽量な手法を提案します。従来のアイデア流暢性や専門家評価といった方法は、議論の微妙なニュアンスを捉えにくく、一貫した実施が困難である一方、大規模言語モデル（LLM）などの計算手法はリソースコストが高い問題があります。本研究では、53件の実際のビジネス会議（平均1.5時間）から得られたデータを用いて、キーワードネットワークの蓄積・非蓄積指標を分析し、主成分分析（PCA）によって主要な特徴を抽出しました。議論の広がり、局所的強度、トピックの多様性の3つの指標を開発し、会議の質を評価しました。これらの指標は、高パフォーマンス会議を効果的に区別し、蓄積的な傾向と瞬間的なダイナミクスの両方を捉えます。本手法は、チームの生産性や協力を向上させるためのスケーラブルかつコスト効率の高い代替手段を提供します。

会議はビジネスの場において、創造性を発揮する、つまりアイデア創出を行う代表的な機会であり、会議における創造性に寄与するような参加者の言動の特徴やスキルを明らかにすることが求められる。発話の分析においては、コーディングスキームを用いた発話の分類(アノテーションの付与)とその結果の頻度分析を行うことで、発話の特徴を定量化することができる。しかし、アノテーションの付与は人力で行うため時間とコストを要するという課題がある。大規模言語モデル(LLM)を使用することで、そういった課題を解決できる可能性があるが、LLMを用いたアノテーションタスクに関して、会議のような口語発話データの分類を行った過去の研究は限られている。本研究では、会議発話データの分析のためのアノテーション付与にLLMの一種であるChatGPTを活用し、精度を保ちつつ効率化できるかを検証することを目的とした。そして、人間2人、人間1人とChatGPT、ChatGPT単体の3パターンのコーディング性能を比較した。その結果、ChatGPTの結果を人間が修正する方法で、精度をある程度保ちつつ作業時間を70%短縮できることが明らかになった。

With AI technology rapidly growing, human creativity is expected to rise. Business meetings foster creativity, which in turn contributes to meeting productivity. However, according to the Harvard Business Review, 71% of meetings are deemed unproductive. Prior to improve creativity in meetings, it is important to examine and define ongoing interactions. Thus, this paper primarily seeks to address this objective (a) clarify meeting phenomena from multiple perspectives, (b) visualize the meeting characteristics based on the analysis results, and (c) identify the rationale behind the distinctive points for future feedback. This paper presents our on-going analysis of the recordings of real business meetings to identify key characteristics that contribute to or are associated with meeting creativity, with the ultimate goal of enhancing meeting outcomes. Behavioral analysis (e.g., the number of utterances, the number of characters) and emotional analyses were employed to uncover the descriptive statistics and distinctive features within the meeting data. Qualitative analysis was conducted to investigate the distinctive features, such as meeting atmosphere and individual traits. We introduced multilevel and user-centric visualization of the analysis result, providing an intuitive understanding of the features for self-evaluation. This approach is expected to provide actionable insights to improve meeting dynamics and overall effectiveness.

During the Covid-19 pandemic, more guidelines were created to teach people how to facilitate meetings online, but few were designed from a cognition-oriented perspective. Additionally, solving complex problems is essential in many occupations. However, the influence of online and face-to-face discussion formats on the performance in complex problem-solving tasks is unclear, even though remote working has become common over the past several few years. Hence, this study aims to answer two research questions: (a) Does problem-solving performance differ between online and face-to-face meetings? and (b) Does facilitation improve problem-solving performance when different formats are used? We conducted experiments with 40 groups using a 2 × 2 factorial design, which were controlled for both facilitation and format. Each group comprised two randomly selected participants, and each problem-solving discussion lasted between 1.5–2 h. The obtained evidence showed that format can influence the performance of balancing intercorrelated factors in a complex scenario, but it does not affect the performance of achieving a predefined goal. Instead, it we found that facilitation is helpful for achieving a predefined goal. Based on the results obtained, we propose future design directions for problem-solving centric computer-supported cooperative work systems from a cognition-oriented perspective.

During the pandemic, people gradually realized the limitations of remote problem-solving collaboration. Thus far, the impact of online platforms on problem-solving cognitive processes has not been thoroughly investigated, despite the active use of such processes in remote work. This study analyzes the differences in cognitive patterns between online and offline problem-solving meetings. Discussion data containing approximately 5,000 utterances were subjected to entropy and content analysis. The results showed a distinct difference in cognitive patterns in discussions conducted on various platforms, as online discussions require more cognitive effort for technological equipment operation. Online meeting solutions were found to be less developed than those generated in offline meetings. According to the results obtained, we propose a series of high-level platform-neutral steps for small-scale problem-solving. The findings not only contribute to building platform-specific facilitation guidelines, but they also aid research on human-computer interaction from the perspective of online discussion platform designs and methods for cognitive process evaluation.

Abstract: Complex problem solving (CPS) has been a field that uses computer-simulated scenarios and has been applied in problem-solving-related studies. However, the problem scenario has not been thoroughly discussed as an essential factor in determining the reliability of the studies. Consequently, there are no systematic principles for scenario design in CPS studies. This study was performed to establish fundamental standards for complexity analysis in the CPS scenario design. We created a high-fidelity problem scenario to investigate the cognitive processes in CPS discussions. The reliability of the system and scenario was validated by five industrial experts. The findings of this study can be applied to future experiment designs, meta-analysis methods, and study replications.

Meetings are one of the most common collaboration formats for complex problem-solving (CPS). This research aims to formulate cognitive-oriented guidelines for productive synchronous CPS discussions. The study proposes a method to analyze the cognitive process and identifies the cognitive process associated with better CPS discussions. A conversation-analysis method was developed. Two indicators—source–outcome retrieval ratio and count of overlapped solution utterances—were proposed to evaluate the CPS discussion’s efficiency and effectiveness. Sixteen experimental CPS discussions were analyzed using this method. Correlation coefficients were applied to ascertain the cognitive features in CPS discussions with different levels of effectiveness and confirmed the applicability and reliability of the proposed methods. The results revealed that a good CPS discussion includes a regular progress summary, discussion conclusion, and high utilization of cognitive sources.

Many human factors researchers have explored the cognitive and behavioral factors that affect team performance through behavioral and verbal protocol analyses. These studies primarily used qualitative analyses of observable behaviors and utterances, which makes it difficult to capture the dynamic and resilient team cooperation process directly. Therefore, it is necessary to develop quantitative indicators or measures to assess dynamic processes in team behavior and communication. Once such appropriate indicators or measures are developed, we can compare the performance of different teams quantitatively and find the features of team cognition that support good performance. In the study of complex problem solving, several studies calculated the entropies of utterances from the results of a qualitative analysis of team communication to detect phase changes in complex problem solving (Wiltshire and Butner, 2017). In addition to entropy, this study calculates the Kullback–Leibler divergence (KL) of utterances in segments for the entire team process to identify dynamic features and irregular segments in team communication. We applied the information theory to quantify the features of utterances in segments for the entire team process to find dynamic features and irregular segments in team communication. We analyzed the utterance data of a three-person team working on a task that required dynamic role assignment and collaboration. We first analyzed the turn-taking and communication contents and then visualized them using recurrence plots to visually find sequential patterns. We then calculated the Kullback–Leibler divergence (KL) and plotted it with sliding windows to analyze the dynamic features in team communication. The results showed that the bias of the content increased with disturbances, which suggests that the proposed indices can be used to capture speech distortions caused by external disturbances.

A complex problem has no clear definition of goals or solutions. Many social challenges fall within the range of complex problems. This research aims to formulate cognition-oriented guidelines for conducting productive CPS discussions. In our previous work, we developed a method to extract the cognitive process of complex problem-solving discussions. In this study, we proposed two indicators, “source-outcome retrieval ratio” and “change in the count of overlapped solution utterances” for the performance evaluation of CPS discussions.

The efficiency of complex problem-solving (CPS) in groups determines the productivity of a society. Current CPS methods are not satisfactory owing to their ineffectiveness and high resource requirements. This research aims to formulate cognition-oriented guidelines for conducting productive CPS discussions. A method for evaluating personal relevancy and perspective toward problem complexity was developed in our previous work to explore the underlying cognitive processes in CPS discussions. This paper presents the updated results, demonstrating that the ability to thoroughly interpret the problem at the beginning of a CPS discussion determines the quality of the discussion.

The efficiency of solving complex problems in groups determines the productivity of a society. Existing guidelines for these collaborations are action-focused, and the few cognitive-oriented ones require time and training to be executed accurately. This research aims to propose intuitive and light-weighted recommendations for Complex-Problem-Solving (CPS) collaborations. The underlying cognitive process in CPS discussion was explored, especially the explicit knowledge used. It was found that episodic memory functions better at expanding the conversation scope, while semantic memory appeared to be a more straightforward foundation to initiate new ideas. Since the episodic memory serves as an outstanding primer in the conversation, the results could imply that the better episodic memory is communicated, the more fluence the discussion could be.