Adrian Groza

Patricia Mateiu, Adrian Groza

We tackle the task of enriching ontologies by automatically translating natural language sentences into Description Logic. Since Large Language Models (LLMs) are the best tools for translations, we fine-tuned a GPT-3 model to convert Natural Language sentences into OWL Functional Syntax. We employ objective and concise examples to fine-tune the model regarding: instances, class subsumption, domain and range of relations, object properties relationships, disjoint classes, complements, cardinality restrictions. The resulted axioms are used to enrich an ontology, in a human supervised manner. The developed tool is publicly provided as a Protge plugin.

Adrian Groza

I shall quantify the logical faults generated by ChatGPT when applied to reasoning tasks. For experiments, I use the 144 puzzles from the library \url{https://users.utcluj.ro/~agroza/puzzles/maloga}~\cite{groza:fol}. The library contains puzzles of various types, including arithmetic puzzles, logical equations, Sudoku-like puzzles, zebra-like puzzles, truth-telling puzzles, grid puzzles, strange numbers, or self-reference puzzles. The correct solutions for these puzzles were checked using the theorem prover Prover9~\cite{mccune2005release} and the finite models finder Mace4~\cite{mccune2003mace4} based on human-modelling in Equational First Order Logic. A first output of this study is the benchmark of 100 logical puzzles. For this dataset ChatGPT provided both correct answer and justification for 7\% only. %, while BARD for 5\%. Since the dataset seems challenging, the researchers are invited to test the dataset on more advanced or tuned models than ChatGPT3.5 with more crafted prompts. A second output is the classification of reasoning faults conveyed by ChatGPT. This classification forms a basis for a taxonomy of reasoning faults generated by large language models. I have identified 67 such logical faults, among which: inconsistencies, implication does not hold, unsupported claim, lack of commonsense, wrong justification. The 100 solutions generated by ChatGPT contain 698 logical faults. That is on average, 7 fallacies for each reasoning task. A third ouput is the annotated answers of the ChatGPT with the corresponding logical faults. Each wrong statement within the ChatGPT answer was manually annotated, aiming to quantify the amount of faulty text generated by the language model. On average, 26.03\% from the generated text was a logical fault.

Adrian Groza, Anca Marginean

We exemplify how Large Language Models are used in both teaching and learning. We also discuss the AI incidents that have already occurred in the education domain, and we argue for the urgent need to introduce AI policies in universities and for the ongoing strategies to regulate AI. Regarding policy for AI, our view is that each institution should have a policy for AI in teaching and learning. This is important from at least twofolds: (i) to raise awareness on the numerous educational tools that can both positively and negatively affect education; (ii) to minimise the risk of AI incidents in education.

Stefan Morar, Adrian Groza, Mihai Pomarlan

We present a method for formalising quantifiers in natural language in the context of human-robot interactions. The solution is based on first-order logic extended with capabilities to represent the cardinality of variables, operating similarly to generalised quantifiers. To demonstrate the method, we designed an end-to-end system able to receive input as natural language, convert it into a formal logical representation, evaluate it, and return a result or send a command to a simulated robot.

Alexandru Motogna, Adrian Groza

An attractive book cover is important for the success of a book. In this paper, we apply Generative Adversarial Networks (GANs) to the book covers domain, using different methods for training in order to obtain better generated images. We interleave GANs with knowledge graphs to alter the input title to obtain multiple possible options for any given title, which are then used as an augmented input to the generator. Finally, we use the discriminator obtained during the training phase to select the best images generated with new titles. Our method performed better at generating book covers than previous attempts, and the knowledge graph gives better options to the book author or editor compared to using GANs alone.

Darius Has, Adrian Groza, Mihai Pomarlan

We propose a solution for handling abort commands given to robots. The solution is exemplified with a running scenario with household kitchen robots. The robot uses planning to find sequences of actions that must be performed in order to gracefully cancel a previously received command. The Planning Domain Definition Language (PDDL) is used to write a domain to model kitchen activities and behaviours, and this domain is enriched with knowledge from online ontologies and knowledge graphs, like DBPedia. We discuss the results obtained in different scenarios.

Marco Pop-Mihali, Adrian Groza

We aim at development white-box machine learning algorithms. We focus here on algorithms for learning axioms in description logic. We extend the Class Expression Learning for Ontology Engineering (CELOE) algorithm contained in the DL-Learner tool. The approach uses multiple search trees and a shared pool of refinements in order to split the search space in smaller subspaces. We introduce the conjunction operation of best class expressions from each tree, keeping the results which give the most information. The aim is to foster exploration from a diverse set of starting classes and to streamline the process of finding class expressions in ontologies. %, particularly in large search spaces. The current implementation and settings indicated that the Forest Mixing approach did not outperform the traditional CELOE. Despite these results, the conceptual proposal brought forward by this approach may stimulate future improvements in class expression finding in ontologies. % and influence. % the way we traverse search spaces in general.

Eduard Popescu, Adrian Groza, Ioana Damian

Diabetic retinopathy is an ocular condition that affects individuals with diabetes mellitus. It is a common complication of diabetes that can impact the eyes and lead to vision loss. One method for diagnosing diabetic retinopathy is the examination of the fundus of the eye. An ophthalmologist examines the back part of the eye, including the retina, optic nerve, and the blood vessels that supply the retina. In the case of diabetic retinopathy, the blood vessels in the retina deteriorate and can lead to bleeding, swelling, and other changes that affect vision. We proposed a method for detecting diabetic diabetic severity levels. First, a set of data-prerpocessing is applied to available data: adaptive equalisation, color normalisation, Gaussian filter, removal of the optic disc and blood vessels. Second, we perform image segmentation for relevant markers and extract features from the fundus images. Third, we apply an ensemble of classifiers and we assess the trust in the system.

Alexandru Lecu, Adrian Groza, Lezan Hawizy

Large language models (LLMs) have significantly advanced the field of natural language generation. However, they frequently generate unverified outputs, which compromises their reliability in critical applications. In this study, we propose an innovative framework that combines structured biomedical knowledge with LLMs through a retrieval-augmented generation technique. Our system develops a thorough knowledge graph by identifying and refining causal relationships and named entities from medical abstracts related to age-related macular degeneration (AMD). Using a vector-based retrieval process and a locally deployed language model, our framework produces responses that are both contextually relevant and verifiable, with direct references to clinical evidence. Experimental results show that this method notably decreases hallucinations, enhances factual precision, and improves the clarity of generated responses, providing a robust solution for advanced biomedical chatbot applications.

Eduard Popescu, Adrian Groza, Andreea Cernat

The task here is to predict the toxicological activity of chemical compounds based on the Tox21 dataset, a benchmark in computational toxicology. After a domain-specific overview of chemical toxicity, we discuss current computational strategies, focusing on machine learning and deep learning. Several architectures are compared in terms of performance, robustness, and interpretability. This research introduces a novel image-based pipeline based on DenseNet121, which processes 2D graphical representations of chemical structures. Additionally, we employ Grad-CAM visualizations, an explainable AI technique, to interpret the model's predictions and highlight molecular regions contributing to toxicity classification. The proposed architecture achieves competitive results compared to traditional models, demonstrating the potential of deep convolutional networks in cheminformatics. Our findings emphasize the value of combining image-based representations with explainable AI methods to improve both predictive accuracy and model transparency in toxicology.

Alexandru-Andrei Avram, Adrian Groza, Alexandru Lecu

The large spread of disinformation across digital platforms creates significant challenges to information integrity. This paper presents a multi-agent system that uses relation extraction to detect disinformation in news articles, focusing on titles and short text snippets. The proposed Agentic AI system combines four agents: (i) a machine learning agent (logistic regression), (ii) a Wikipedia knowledge check agent (which relies on named entity recognition), (iii) a coherence detection agent (using LLM prompt engineering), and (iv) a web-scraped data analyzer that extracts relational triplets for fact checking. The system is orchestrated via the Model Context Protocol (MCP), offering shared context and live learning across components. Results demonstrate that the multi-agent ensemble achieves 95.3% accuracy with an F1 score of 0.964, significantly outperforming individual agents and traditional approaches. The weighted aggregation method, mathematically derived from individual agent misclassification rates, proves superior to algorithmic threshold optimization. The modular architecture makes the system easily scalable, while also maintaining details of the decision processes.

Mircea-Serban Vasiliniuc, Adrian Groza

The aim of this study is to evaluate the performance of AI-assisted programming in actual mobile development teams that are focused on native mobile languages like Kotlin and Swift. The extensive case study involves 16 participants and 2 technical reviewers, from a software development department designed to understand the impact of using LLMs trained for code generation in specific phases of the team, more specifically, technical onboarding and technical stack switch. The study uses technical problems dedicated to each phase and requests solutions from the participants with and without using AI-Code generators. It measures time, correctness, and technical integration using ReviewerScore, a metric specific to the paper and extracted from actual industry standards, the code reviewers of merge requests. The output is converted and analyzed together with feedback from the participants in an attempt to determine if using AI-assisted programming tools will have an impact on getting developers onboard in a project or helping them with a smooth transition between the two native development environments of mobile development, Android and iOS. The study was performed between May and June 2023 with members of the mobile department of a software development company based in Cluj-Napoca, with Romanian ownership and management.

Adrian Groza

The proposed approach is to formalise the probabilistic puzzle in equational FOL. Two formalisations are needed: one theory for all models of the given puzzle, and a second theory for the favorable models. Then Mace4 - that computes all the interpretation models of a FOL theory - is called twice. First, it is asked to compute all the possible models M p .Second, the additional constraint is added, and Mace4 computes only favourabile models M f. Finally, the definition of probability is applied: the number of favorable models is divided by the number of possible models. The proposed approach equips students from the logic tribe to find the correct solution for puzzles from the probabilitistic tribe, by using their favourite instruments: modelling and formalisation. I have exemplified here five probabilistic puzzles and how they can be solved by translating the min FOL and then find the corresponding interpretation models. Mace4 was the tool of choice here. Ongoing work is investigating the limits of this method on various collections of probabilistic puzzles

Roxana Szomiu, Adrian Groza

We provide here a dataset for tasks related to natural language understanding and natural language inference. The dataset contains logical puzzles in natural language from three domains: comparing puzzles, knighs and knaves, and zebra puzzles. Each puzzle is associated with the entire set of atomic questions that can be generated based on the relations and individuals occurring in the text. For each question we provide the correct answer: entailment, contradiction or ambiguity. The answer's correctness is verified against theorem provers. Good puzzles have two properties: (i) each piece of information is necessary and (ii) no unnecessary information is provided. These properties make puzzles interesting candidates for machine comprehension tasks.

Adrian Groza, Cristian Nitu

We developed a system able to automatically solve logical puzzles in natural language. Our solution is composed by a parser and an inference module. The parser translates the text into first order logic (FOL), while the MACE4 model finder is used to compute the models of the given FOL theory. We also empower our software agent with the capability to provide Yes/No answers to natural language questions related to each puzzle. Moreover, in line with Explainalbe Artificial Intelligence (XAI), the agent can back its answer, providing a graphical representation of the proof. The advantage of using reasoning for Natural Language Understanding (NLU) instead of Machine learning is that the user can obtain an explanation of the reasoning chain. We illustrate how the system performs on various types of natural language puzzles, including 382 knights and knaves puzzles. These features together with the overall performance rate of 80.89\% makes the proposed solution an improvement upon similar solvers for natural language understanding in the puzzles domain.

Adrian Groza

In the context of the Covid-19 pandemic, many were quick to spread deceptive information. I investigate here how reasoning in Description Logics (DLs) can detect inconsistencies between trusted medical sources and not trusted ones. The not-trusted information comes in natural language (e.g. "Covid-19 affects only the elderly"). To automatically convert into DLs, I used the FRED converter. Reasoning in Description Logics is then performed with the Racer tool.

Adrian Groza, Madalina Mand Nagy

Parkinson's disease is the second most common neurodegenerative disease, affecting more than 1.2 million people in Europe. Medications are available for the management of its symptoms, but the exact cause of the disease is unknown and there is currently no cure on the market. To better understand the relations between new findings and current medical knowledge, we need tools able to analyse published medical papers based on natural language processing and tools capable to identify various relationships of new findings with the current medical knowledge. Our work aims to fill the above technological gap. To identify conflicting information in medical documents, we enact textual entailment technology. To encapsulate existing medical knowledge, we rely on ontologies. To connect the formal axioms in ontologies with natural text in medical articles, we exploit ontology verbalisation techniques. To assess the level of disagreement between human agents with respect to a medical issue, we rely on fuzzy aggregation. To harmonize this disagreement, we design mediation protocols within a multi-agent framework.

Adrian Groza, Oana Popa

In the medical domain, the continuous stream of scientific research contains contradictory results supported by arguments and counter-arguments. As medical expertise occurs at different levels, part of the human agents have difficulties to face the huge amount of studies, but also to understand the reasons and pieces of evidences claimed by the proponents and the opponents of the debated topic. To better understand the supporting arguments for new findings related to current state of the art in the medical domain we need tools able to identify arguments in scientific papers. Our work here aims to fill the above technological gap. Quite aware of the difficulty of this task, we embark to this road by relying on the well-known interleaving of domain knowledge with natural language processing. To formalise the existing medical knowledge, we rely on ontologies. To structure the argumentation model we use also the expressivity and reasoning capabilities of Description Logics. To perform argumentation mining we formalise various linguistic patterns in a rule-based language. We tested our solution against a corpus of scientific papers related to breast cancer. The run experiments show a F-measure between 0.71 and 0.86 for identifying conclusions of an argument and between 0.65 and 0.86 for identifying premises of an argument.

Adrian Groza, Calin Cara, Sergiu Zaporojan et al.

A warning system for assisting drivers during overtaking maneuvers is proposed. The system relies on Car-2-Car communication technologies and multi-agent systems. A protocol for safety overtaking is proposed based on ACL communicative acts. The mathematical model for safety overtaking used Kalman filter to minimize localization error.

Adrian Groza, Lidia Corde

Our aim is to extract information about literary characters in unstructured texts. We employ natural language processing and reasoning on domain ontologies. The first task is to identify the main characters and the parts of the story where these characters are described or act. We illustrate the system in a scenario in the folktale domain. The system relies on a folktale ontology that we have developed based on Propp's model for folktales morphology.

Adrian Groza, Roxana Szabo

We propose a system for automated essay grading using ontologies and textual entailment. The process of textual entailment is guided by hypotheses, which are extracted from a domain ontology. Textual entailment checks if the truth of the hypothesis follows from a given text. We enact textual entailment to compare students answer to a model answer obtained from ontology. We validated the solution against various essays written by students in the chemistry domain.

Adrian Groza

Small wind projects encounter difficulties to be efficiently deployed, partly because wrong way data and information are managed. Ontologies can overcome the drawbacks of partially available, noisy, inconsistent, and heterogeneous data sources, by providing a semantic middleware between low level data and more general knowledge. In this paper, we engineer an ontology for the wind energy domain using description logic as technical instrumentation. We aim to integrate corpus of heterogeneous knowledge, both digital and human, in order to help the interested user to speed-up the initialization of a small-scale wind project. We exemplify one use case scenario of our ontology, that consists of automatically checking whether a planned wind project is compliant or not with the active regulations.

Ioan Alfred Letia, Adrian Groza

When perturbation or unexpected events do occur, agents need protocols for repairing or reforming the supply chain. Unfortunate contingency could increase too much the cost of performance, while breaching the current contract may be more efficient. In our framework the principles of contract law are applied to set penalties: expectation damages, opportunity cost, reliance damages, and party design remedies, and they are introduced in the task dependency model

Adrian Groza, Sergiu Indrie

This research advocates the idea of combining argumentation theory with the social web technology, aiming to enact large scale or mass argumentation. The proposed framework allows mass-collaborative editing of structured arguments in the style of semantic wikipedia. The long term goal is to apply the abstract machinery of argumentation theory to more practical applications based on human generated arguments, such as deliberative democracy, business negotiation, or self-care. The ARGNET system was developed based on ther Semantic MediaWiki framework and on the Argument Interchange Format (AIF) ontology.

Adrian Groza, Gabriel Barbur, Bogdan Blaga

This paper presents the OntoRich framework, a support tool for semi-automatic ontology enrichment and evaluation. The WordNet is used to extract candidates for dynamic ontology enrichment from RSS streams. With the integration of OpenNLP the system gains access to syntactic analysis of the RSS news. The enriched ontologies are evaluated against several qualitative metrics.

Ioan Alfred Letia, Adrian Groza

In the interaction between agents we can have an explicative discourse, when communicating preferences or intentions, and a normative discourse, when considering normative knowledge. For justifying their actions our agents are endowed with a Justification and Explanation Logic (JEL), capable to cover both the justification for their commitments and explanations why they had to act in that way, due to the current situation in the environment. Social commitments are used to formalise justificatory and explanatory patterns. The combination of ex- planation, justification, and commitments