The role of knowledge graphs in chatbots
Bibliographic record
Abstract
Purpose This study aims to investigate the applications of knowledge graphs in developing artificial intelligence (AI) assistants and chatbots by reviewing scholarly publications from different lenses and dimensions. The authors also analyze the various AI approaches used for knowledge graph-driven chatbots and discuss how implementing these techniques makes a difference in technology. Design/methodology/approach Over recent years, chatbots have emerged as a transformational force in interacting with the digital world in various domains, including customer service and personal assistants. Recently, chatbots have become a revolutionary tool for interacting with the digital world in various contexts, such as personal assistants and customer support. Incorporating knowledge graphs considerably improved the capabilities of chatbots by allowing them access to massive knowledge bases and enhancing their ability to understand queries. Furthermore, knowledge graphs enable chatbots to understand semantic links between elements and improve response quality. This study highlights the role of knowledge graphs in chatbots following a systematic review approach. They have been integrated into major health-care, education and business domains. Beyond improving information retrieval, knowledge graphs enhance the user experience and increase the range of fields in which chatbots can be used. Improving and enriching chatbot answers was also identified as one of the main advantages of knowledge graphs. This enriched response can increase user confidence and improve the accuracy of chatbot interactions, making them more trustworthy information sources. Findings Knowledge graph-based chatbots leverage extensive data retrieval to provide accurate and enriched responses, increasing user confidence and experience without requiring extensive training. The three major domains where knowledge graph-based chatbots have been used are health care, education and business. Practical implications Knowledge graph-based chatbots can better comprehend user queries and respond with relevant information efficiently without extensive training. Furthermore, knowledge graphs enable chatbots to understand semantic links between elements, allowing them to answer complicated and multi-faceted questions. This semantic comprehension improves response quality, making chatbots more successful in providing accurate and valuable information in various domains. Furthermore, knowledge graphs enable chatbots to provide consumers with individualized experiences by storing and recalling individual preferences, history or previous encounters. This study analyzes the role of knowledge graphs in chatbots following a systematic review approach. This study reviewed state-of-the-art articles to understand where and how chatbots have used knowledge graphs. The authors found health care, business and education as three main areas in which knowledge-graph-based chatbots have been mostly used. Chatbots have been developed in text, voice and visuals using various machine learning models, particularly natural language pocessing, to develop recommender systems to recommend suitable items, content or services based on user preferences and item associations. Originality/value This paper provides a comprehensive review of the current state of the field in using knowledge graphs in chatbots, focusing on machine learning models, domains and communication channels. The study highlights the prevalence of text and voice channels over visual ones and identifies research gaps and future directions. The paper’s insights can inform the design and development of chatbots using knowledge graphs and benefit both researchers and practitioners in AI, natural language processing and human–computer interaction. The paper is of interest to professionals in domains such as health care, education and business.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.068 | 0.192 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.018 | 0.011 |
| Science and technology studies | 0.002 | 0.008 |
| Scholarly communication | 0.010 | 0.015 |
| Open science | 0.002 | 0.006 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".