Martin, Andreas
Lade...
E-Mail-Adresse
Geburtsdatum
Projekt
Organisationseinheiten
Berufsbeschreibung
Nachname
Martin
Vorname
Andreas
Name
Martin, Andreas
44 Ergebnisse
Suchergebnisse
Gerade angezeigt 1 - 10 von 44
Publikation An LLM-aided Enterprise Knowledge Graph (EKG) engineering process(Stanford University, 2024) Laurenzi, Emanuele; Mathys, Adrian; Martin, Andreas; Petrick, Ron; Geib, ChristopherConventional knowledge engineering approaches aiming to create Enterprise Knowledge Graphs (EKG) still require a high level of manual effort and high ontology expertise, which hinder their adoption across industries. To tackle this issue, we explored the use of Large Language Models (LLMs) for the creation of EKGs through the lens of a design-science approach. Findings from the literature and from expert interviews led to the creation of the proposed artefact, which takes the form of a six-step process for EKG development. Scenarios on how to use LLMs are proposed and implemented for each of the six steps. The process is then evaluated with an anonymised data set from a large Swiss company. Results demonstrate that LLMs can support the creation of EKGs, offering themselves as a new aid for knowledge engineers.04B - Beitrag KonferenzschriftPublikation Domain-specific embeddings for question-answering systems. FAQs for health coaching(Stanford University, 2024) Martin, Andreas; Pande, Charuta; Schwander, Sandro; Ajuwon, Ademola J.; Pimmer, Christoph; Petrick, Ron; Geib, ChristopherFAQs are widely used to respond to users’ knowledge needs within knowledge domains. While LLM might be a promising way to address user questions, they are still prone to hallucinations i.e., inaccurate or wrong responses, which, can, inter alia, lead to massive problems, including, but not limited to, ethical issues. As a part of the healthcare coach chatbot for young Nigerian HIV clients, the need to meet their information needs through FAQs is one of the main coaching requirements. In this paper, we explore if domain knowledge in HIV FAQs can be represented as text embeddings to retrieve similar questions matching user queries, thus improving the understanding of the chatbot and the satisfaction of the users. Specifically, we describe our approach to developing an FAQ chatbot for the domain of HIV. We used a predefined FAQ question-answer knowledge base in English and Pidgin co-created by HIV clients and experts from Nigeria and Switzerland. The results of the post-engagement survey show that the chatbot mostly understood the user’s questions and could identify relevant matching questions and retrieve an appropriate response.04B - Beitrag KonferenzschriftPublikation Semantic verification in large language model-based retrieval augmented generation(AAAI Press, 2024) Martin, Andreas; Witschel, Hans Friedrich; Mandl, Maximilian; Stockhecke, Mona; Petrick, Ron; Geib, Christopher04B - Beitrag KonferenzschriftPublikation ChEdBot: designing a domain-specific conversational agent in a simulational learning environment using LLMs(AAAI Press, 2024) Martin, Andreas; Pande, Charuta; Witschel, Hans Friedrich; Mathez, Judith; Petrick, Ron; Geib, Christopher04B - Beitrag KonferenzschriftPublikation AAAI-MAKE 2023: Challenges requiring the combination of machine learning and knowledge engineering(Association for the Advancement of Artificial Intelligence, 2023) Martin, Andreas01A - Beitrag in wissenschaftlicher ZeitschriftPublikation Ontology-driven enhancement of process mining with domain knowledge(Sun SITE, Informatik V, RWTH Aachen, 2023) Eichele, Simon; Hinkelmann, Knut; Spahic, Maja; Martin, Andreas; Fill, Hans-Georg; Gerber, Aurona; Hinkelmann, Knut; Lenat, Doug; Stolle, Reinhard; Harmelen, Frank vanProcess mining is a technique used to analyze and understand business processes. It uses as input the event log, a type of data used to represent the sequence of activities occurring within a business process. An event log typically contains information such as the case ID, the performed activity’s name, the activity’s timestamp, and other data associated with the activity. By analyzing event logs, organizations can gain a deeper understanding of their business processes, identify areas for improvement, and make data-driven decisions to optimize their operations. However, as the event logs contain data collected from different systems involved in the process, such as ERP, CRM, or WfMS systems, they often lack the necessary context and knowledge to analyze and fully comprehend business processes. By extending the event logs with domain knowledge, organizations can gain a more complete and accurate insight into their business processes and make more informed decisions about optimizing them. This paper presents an approach for enhancing process mining with domain knowledge preserved in domain-specific OWL ontologies. Event logs are typically stored in structured form in relational databases. This approach first converts the process data into an event log which is then mapped with ontology concepts. The ontology contains classes and individuals representing background knowledge of the domain, which supports the understanding of the data. A class for the specific activities forms the link between the event log and the ontology. In this manner, it is possible to map the domain knowledge to a particular case and activity. This allows to determine conditions that must be satisfied for executing tasks and to prune discovered process models if they are too complex. This approach is demonstrated using data from the student admission process at FHNW and has been implemented in Protégé.04B - Beitrag KonferenzschriftPublikation Visualisierung von Mustern für hybrides Lernen und Reasoning mit menschlicher Beteiligung(Springer, 2023) Witschel, Hans Friedrich; Pande, Charuta; Martin, Andreas; Laurenzi, Emanuele; Hinkelmann, Knut; Dornberger, Rolf04A - Beitrag SammelbandPublikation Towards hybrid dialog management strategies for a health coach chatbot(2023) Pande, Charuta; Martin, Andreas; Pimmer, ChristophWe present an iterative and incremental approach to designing dialog management for a health coach chatbot based on our in-progress research. The requirements are derived from the coaching needs of young people living with HIV. We identify a hybrid dialog management approach to address different coaching needs as well as dialog acts to enable smooth conversations. In addition, relevant technical components were identified to be integrated into the dialogs to improve user experience.04B - Beitrag KonferenzschriftPublikation A flexible, extendable and adaptable model to support AI coaching(Springer, 2023) Duhan, Ritu; Pande, Charuta; Martin, Andreas; Hinkelmann, Knut; López-Pellicer, Francisco J.; Polini, AndreaWe present a model based on coaching definitions, concepts, and theories to support AI coaching. The model represents the evidence-based coaching practice in different coaching domains by identifying the common elements in the coaching process. We then map the elements of the coaching model with Conversational AI design and development strategies to highlight how an AI coach can be instantiated from the model. We showcase the instantiation through an example use case of an HIV coaching chatbot.04B - Beitrag KonferenzschriftPublikation A hybrid intelligent approach combining machine learning and a knowledge graph to support academic journal publishers addressing the Reviewer Assignment Problem (RAP)(Sun SITE, Informatik V, RWTH Aachen, 2023) Rordorf, Dietrich Hans-Paul; Käser, Josua; Crego Corot, Alfredo Etienne; Laurenzi, Emanuele; Martin, Andreas; Fill, Hans-Georg; Gerber, Aurona; Hinkelmann, Knut; Lenat, Doug; Stolle, Reinhard; Harmelen, FrankThis paper presents a hybrid intelligent approach that combines natural language processing (NLP) and knowledge engineering to address the Reviewer Assignment Problem (RAP) in scientific peer-review. The approach uses NLP techniques to match a new document with subject experts, and it employs a knowledge graph to identify conflicts of interest (COIs) between the authors of a document and potential reviewers. The approach detects three types of COIs: direct co-authorship, second-level coauthorship, and collaborators from the same institutions. Further, it uses semantic text similarity (STS) matching for peer-reviewing of documents in journals, where potential reviewers are screened from large literature databases. The research approach follows the Design Science Research methodology, where a prototypical system is designed based on the requirements elicited from both the literature and from primary data collection conducted in a publishing house. The approach is evaluated by implementing real-world use cases in the working prototype and by conducting a focus group with potential users, i.e., editors. © 2023 CEUR-WS. All rights reserved.04B - Beitrag Konferenzschrift