Shinobu Hasegawa’s research while affiliated with Japan Advanced Institute of Science and Technology and other places

In recent years, researchers have witnessed rapid advancements in conducting paper surveys using generative AI, enhancing survey efficiency to some extent. However, today's generative AI lacks deep research training to analyze logical threads woven across multiple papers. A concise visualization method is also expected to present logical connections among various papers. These logical threads are often implicit in the issue ontology authors commonly employ when writing papers. Building on this issue ontology, our method utilizes Dynamic Programming with multiple agents to generate an insight path. The key feature of this approach is the collaboration of multiple agents to adapt to a complex environment and make optimized decisions on issue ontology selection. This path aims to succinctly express longitudinal logical connections among multiple papers, including commonality, difference, and inheritance.

In the flourishing era of information science, effective comprehension, observation, and insight from various academic papers are crucial skills for researchers. However, this can be challenging for beginners without enough research training. The current knowledge graphs and automatic summarization systems used in research insight surveys rarely highlight the similarities and differences among multiple papers based on agreed-upon expert features. This can make novice researchers difficult to understand the logical connections between research concepts. Therefore, this study is committed to assisting researchers in conducting Cross-sectional Insight Survey. It offers a concise diff-table output format, tailored from the perspective of expert consensus. This study aims to generating abstractive summarization based on the viewpoints of expert consensus and showing the differences under these consensus. The final output is in the form of a concise diff-table to assist researchers in conducting Cross-sectional Insight Survey. Our evaluation demonstrates that our generated diff-table outperforms the baseline in terms of BERTScore and conciseness.

RQ : How to help researchers gain effective insights by identifying connections between the evolution and relevance across multiple research papers?

With the rapid development of the internet in the past decade, it has become increasingly important to extract valuable information from vast resources efficiently, which is crucial for establishing a comprehensive digital ecosystem, particularly in the context of research surveys and comprehension. The foundation of these tasks focuses on accurate extraction and deep mining of data from scientific documents, which are essential for building a robust data infrastructure. However, parsing raw data or extracting data from complex scientific documents have been ongoing challenges. Current data extraction methods for scientific documents typically use rule-based (RB) or machine learning (ML) approaches. However, using rule-based methods can incur high coding costs for articles with intricate typesetting. Conversely, relying solely on machine learning methods necessitates annotation work for complex content types within the scientific document, which can be costly. Additionally, few studies have thoroughly defined and explored the hierarchical layout within scientific documents. The lack of a comprehensive definition of the internal structure and elements of the documents indirectly impacts the accuracy of text classification and object recognition tasks. From the perspective of analyzing the standard layout and typesetting used in the specified publication, we propose a new document layout analysis framework called Compartment and Text Blocks Refinement (CTBR). Firstly, we define scientific documents into hierarchical divisions: base domain, compartment, and text blocks. Next, we conduct an in-depth exploration and classification of the meanings of text blocks. Finally, we utilize the results of text block classification to implement object recognition within scientific documents based on rule-based compartment segmentation. For the experiment, we used the well-known ACL format proceeding articles as experimental data for the validation experiment. The experiment shows that our approach achieved over 95% text block classification accuracy and 90% object recognition accuracy for tables and figures.

With the exponential growth in the number of papers and the trend of AI research, the use of Generative AI for information retrieval and question-answering has become popular for conducting research surveys. However, novice researchers unfamiliar with a particular field may not significantly improve their efficiency in interacting with Generative AI because they have not developed divergent thinking in that field. This study aims to develop an in-depth Survey Forest Diagram that guides novice researchers in divergent thinking about the research topic by indicating the citation clues among multiple papers, to help expand the survey perspective for novice researchers.

Disengagement of students during online learning significantly impacts the effectiveness of online education. Thus, accurately estimating when students are not engaged is a critical aspect of online-learning research. However, the inherent characteristics of public datasets often lead to issues of class imbalances and data insufficiency. Moreover, the instability of video time-series data further complicates data processing in related research. Our research aims to tackle class imbalances and instability of video time-series data in estimating learner engagement, particularly in scenarios with limited data. In the present paper, we introduce “Skipped Moving Average”, an innovative oversampling technique designed to augment video time-series data representing disengaged students. Furthermore, we employ long short-term memory (LSTM) and long short-term memory fully convolutional network (LSTM-FCN) models to evaluate the effectiveness of our method and compare it to the synthetic minority over-sampling technique (SMOTE). This approach ensures a thorough evaluation of our method’s effectiveness in addressing video time-series data imbalances and in enhancing the accuracy of engagement estimation. The results demonstrate that our proposed method outperforms others in terms of both performance and stability across sequence deep learning models.

Novice researchers often face difficulties in understanding a multitude of academic papers and grasping the fundamentals of a new research field. To solve such problems, the knowledge graph supporting research survey is gradually being developed. Existing keyword-based knowledge graphs make it difficult for researchers to deeply understand abstract concepts. Meanwhile, novice researchers may find it difficult to use ChatGPT effectively for research surveys due to their limited understanding of the research field. Without the ability to ask proficient questions that align with key concepts, obtaining desired and accurate answers from this large language model (LLM) could be inefficient. This study aims to help novice researchers by providing a fish-bone diagram that includes causal relationships, offering an overview of the research topic. The diagram is constructed using the issue ontology from academic papers, and it offers a broad, highly generalized perspective of the research field, based on relevance and logical factors. Furthermore, we evaluate the strengths and improvable points of the fish-bone diagram derived from this study's development pattern, emphasizing its potential as a viable tool for supporting research survey.

In academic presentation, the structure design of presentation is critical for making the presentation logical and understandable. However, it is difficult for novice researchers to construct required academic presentation structure due to the flexibility in structure creation. To help novice researchers revise and improve their presentation structure, we propose an academic presentation structure modification support system based on structural elements of the presentation slides. In the proposed system, we build a presentation structural elements model (PSEM) that represents the essential structural elements and their relations to clarify the ideal structure of academic presentation. Based on the PSEM, we also designed two evaluation indices to evaluate the academic presentation structure. To evaluate the proposed system with real-world data, we construct a web application that generates evaluation and feedback to academic presentation slides. The experimental results demonstrate the effectiveness of the proposed system.

Research surveys have always posed a challenge for novice researchers who lack of research training. These researchers struggle to understand the directions within their research topic, and the discovery of new research findings within a short time. One way to provide intuitive assistance to novice researchers is by o↵ering relevant knowledge graphs(KG) and recommending related academic papers. However, existing navigation knowledge graphs primarily rely on keywords in the research field and often fail to present the logical hierarchy among multiple related papers clearly. Moreover, most recommendation systems for academic papers simply rely on high text similarity, which can leave researchers confused as to why a particular article is being recommended. They may lack of grasp important information about the insight connection between "Issue resolved" and "Issue finding" that they hope to obtain. To address these issues, this study aims to support research insight surveys for beginner researchers by establishing a hierarchical tree-structured knowledge graph that reflects the inheritance insight of research topics and the relevance insight among the academic papers.