Named entity recognition allows the automated extraction of structured domain-related information from unstructured textual data. Our study explores the task of ontology-driven entity recognition, a sequence labelling process for custom named entity recognition for the domain of dementia, specifically from unstructured forum texts where unprofessional caregivers of people with dementia discuss the challenges they face related to agitation. The targeted corpus is loosely structured, contains ambiguous sentences and vocabulary that does not match the agitation-related medical vocabulary. To address the above challenges, we propose a pipeline that involves the following steps: 1) development of an annotation codebook; 2) annotation of a textual corpus collected from dementia forums, consisting of 45,216 sentences (775 questions and 5571 answers); 3) data augmentation to reduce the imbalance in the corpus; 4) training of a bidirectional LSTM model and a transformer model; 5) comparison of the results with those from few shot- and zero-shot based prompt engineering techniques using a pretrained large language model (LLaMa 3). The results showed that LLaMa 3 was more robust than traditional neural networks and transformer models in detecting underrepresented entities. Furthermore, the study demonstrates that data augmentation improves the entity recognition task when fine-tuning deep learning models. The paper illustrates the challenges of ontology-driven entity recognition in real-world datasets and proposes a roadmap to addressing them that is potentially transferable to other real-world domains.

This study provides a systematic evaluation of prompt engineering strategies for Named Entity Recognition in Nepali, a low-resource language with high similarity to Hindi, by leveraging Hindi-capable Meta’s LLaMA 3.3:70B model. Four prompting techniques—Baseline, Chain-of-Thought, Self-Refine, and Least-toMost—are assessed in both zero-shot and fewshot settings. As a novel contribution, we propose an entity-aware sentence selection strategy that prioritizes example diversity and entity coverage for few-shot prompting. Experimental results show that, without Nepali examples, zero-shot and one-shot prompts frequently yield unstructured or hallucinated outputs, underscoring the limitations of cross-lingual capabilities without in-context supervision. However, including even a small number of carefully selected Nepali examples—sometimes as few as ten—substantially enhances model performance, with the Least-to-Most approach achieving the highest F1 scores. These findings highlight the potential of prompt-based adaptation and principled example curation for extending LLM capabilities to related, low-resource languages, offering a practical alternative to full model fine-tuning.

This study investigates the potential of cross-lingual transfer learning for Named Entity Recognition (NER) between Hindi and Nepali, two languages that, despite their linguistic similarities, face significant disparities in available resources. By leveraging multilingual BERT models, including RemBERT, BERT Multilingual, MuRIL, and DistilBERT Multilingual, the research examines whether pre-training them on a resource-rich language like Hindi can enhance NER performance in a resource-constrained language like Nepali and vice versa. The study conducts experiments in both monolingual and cross-lingual settings to evaluate the models’ effectiveness in transferring linguistic knowledge between the two languages. The findings reveal that while RemBERT and MuRIL perform well in monolingual contexts—RemBERT excelling in Hindi and MuRIL in Nepali—BERT Multilingual performs comparatively best in cross-lingual scenarios, in generalizing features across the languages. Although DistilBERT Multilingual demonstrates slightly lower performance in cross-lingual tasks, it balances efficiency with competitive results. The study underscores the importance of model selection based on linguistic and resource-specific contexts, highlighting that general-purpose models like BERT Multilingual are particularly well-suited for cross-lingual applications.

Recent attempts at behaviour understanding through language grounding have shown that it is possible to automatically generate models for planning problems from textual instructions. One drawback of these approaches is that they either do not make use of the semantic structure behind the model elements identified in the text, or they manually incorporate a collection of concepts with semantic relationships between them. We call this collection of knowledge situation model. The situation model introduces additional context information to the model. It could also potentially reduce the complexity of the planning problem compared to models that do not use situation models. To address this problem, we propose an approach that automatically generates the situation model from textual instructions. The approach is able to identify various hierarchical, spatial, directional, and causal relations. We use the situation model to automatically generate planning problems in a PDDL notation and we show that the situation model reduces the complexity of the PDDL model in terms of number of operators and branching factor compared to planning models that do not make use of situation models.

Different approaches for behaviour understanding rely on textual instructions to generate models of human behaviour. These approaches usually use state of the art parsers to obtain the part of speech (POS) meaning and dependencies of the words in the instructions. For them it is essential that the parser is able to correctly annotate the instructions and especially the verbs as they describe the actions of the person. State of the art parsers usually make errors when annotating textual instructions, as they have short sentence structure often in imperative form. The inability of the parser to identify the verbs results in the inability of behaviour understanding systems to identify the relevant actions. To address this problem, we propose a simple rule-based model that attempts to correct any incorrectly annotated verbs. We argue that the model is able to significantly improve the parser’s performance without the need of additional training data. We evaluate our approach by extracting the actions from 61 textual instructions annotated only with the Stanford parser and once again after applying our model. The results show a significant improvement in the recognition rate when applying the rules (75% accuracy compared to 68% without the rules, p-value < 0.001).