arXiv subjects

Beatrice Alex

Publications and source records attributed to Beatrice Alex.

At least 19 recordsLinked to original sources

arXiv2026

Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads

In long-context use, large language models frequently synthesize answers from the meaning of a relevant context span rather than literally copy-pasting them. Identifying which attention heads perform this synthesis matters for interpreting long-context model behavior. Yet existing detectors miss these heads by construction: they reward heads whose attended token matches the generated token, a literal-copy criterion that captures where a head reads but not what it writes through its output-value (OV) circuit, the very mechanism that carries non-literal retrieval. We introduce Logit-Contribution Scoring (LOCOS), a write-aware detector that scores each head by the projection of its OV-circuit output onto the answer-token unembedding direction, contrasting needle and off-needle source positions in a single forward pass. Across three model families (Qwen3, Gemma-3, OLMo-3.1), mean-ablating the top LOCOS heads on the NoLiMa non-literal retrieval benchmark collapses ROUGE-L at lower head counts than prior attention-based detections; on Qwen3-8B, ablating 50 heads drives ROUGE-L from 0.401 to 0.000 while the strongest baseline still retains 0.292. The selected heads are retrieval-specific: parametric recall and arithmetic reasoning stay at baseline under the same ablation. On Qwen3-8B, the same ablation also drops MuSiQue from 0.55 to 0.08 and BABI-Long from 0.62 to 0.20, while a random-heads control stays within 0.05 of baseline.

cs.CL↗

arXiv2026

GaelEval: Benchmarking LLM Performance for Scottish Gaelic

Multilingual large language models (LLMs) often exhibit emergent 'shadow' capabilities in languages without official support, yet their performance on these languages remains uneven and under-measured. This is particularly acute for morphosyntactically rich minority languages such as Scottish Gaelic, where translation benchmarks fail to capture structural competence. We introduce GaelEval, the first multi-dimensional benchmark for Gaelic, comprising: (i) an expert-authored morphosyntactic MCQA task; (ii) a culturally grounded translation benchmark and (iii) a large-scale cultural knowledge Q&A task. Evaluating 19 LLMs against a fluent-speaker human baseline ($n=30$), we find that Gemini 3 Pro Preview achieves $83.3\%$ accuracy on the linguistic task, surpassing the human baseline ($78.1\%$). Proprietary models consistently outperform open-weight systems, and in-language (Gaelic) prompting yields a small but stable advantage (+$2.4\%$). On the cultural task, leading models exceed $90\%$ accuracy, though most systems perform worse under Gaelic prompting and absolute scores are inflated relative to the manual benchmark. Overall, GaelEval reveals that frontier models achieve above-human performance on several dimensions of Gaelic grammar, demonstrates the effect of Gaelic prompting and shows a consistent performance gap favouring proprietary over open-weight models.

cs.CL↗

arXiv2026

GS-BrainText: A Multi-Site Brain Imaging Report Dataset from Generation Scotland for Clinical Natural Language Processing Development and Validation

We present GS-BrainText, a curated dataset of 8,511 brain radiology reports from the Generation Scotland cohort, of which 2,431 are annotated for 24 brain disease phenotypes. This multi-site dataset spans five Scottish NHS health boards and includes broad age representation (mean age 58, median age 53), making it uniquely valuable for developing and evaluating generalisable clinical natural language processing (NLP) algorithms and tools. Expert annotations were performed by a multidisciplinary clinical team using an annotation schema, with 10-100% double annotation per NHS health board and rigorous quality assurance. Benchmark evaluation using EdIE-R, an existing rule-based NLP system developed in conjunction with the annotation schema, revealed some performance variation across health boards (F1: 86.13-98.13), phenotypes (F1: 22.22-100) and age groups (F1: 87.01-98.13), highlighting critical challenges in generalisation of NLP tools. The GS-BrainText dataset addresses a significant gap in available UK clinical text resources and provides a valuable resource for the study of linguistic variation, diagnostic uncertainty expression and the impact of data characteristics on NLP system performance.

cs.CL↗

arXiv2025

Inverse Scaling in Test-Time Compute

We construct evaluation tasks where extending the reasoning length of Large Reasoning Models (LRMs) deteriorates performance, exhibiting an inverse scaling relationship between test-time compute and accuracy. Our evaluation tasks span four categories: simple counting tasks with distractors, regression tasks with spurious features, deduction tasks with constraint tracking, and advanced AI risks. We identify five distinct failure modes when models reason for longer: 1) Claude models become increasingly distracted by irrelevant information; 2) OpenAI o-series models resist distractors but overfit to problem framings; 3) models shift from reasonable priors to spurious correlations; 4) all models show difficulties in maintaining focus on complex deductive tasks; and 5) extended reasoning may amplify concerning behaviors, with Claude Sonnet 4 showing increased expressions of self-preservation. These findings suggest that while test-time compute scaling remains promising for improving model capabilities, it may inadvertently reinforce problematic reasoning patterns. Our results demonstrate the importance of evaluating models across diverse reasoning lengths to identify and address these failure modes in LRMs.

cs.AI↗

arXiv2025

Adverse Event Extraction from Discharge Summaries: A New Dataset, Annotation Scheme, and Initial Findings

In this work, we present a manually annotated corpus for Adverse Event (AE) extraction from discharge summaries of elderly patients, a population often underrepresented in clinical NLP resources. The dataset includes 14 clinically significant AEs-such as falls, delirium, and intracranial haemorrhage, along with contextual attributes like negation, diagnosis type, and in-hospital occurrence. Uniquely, the annotation schema supports both discontinuous and overlapping entities, addressing challenges rarely tackled in prior work. We evaluate multiple models using FlairNLP across three annotation granularities: fine-grained, coarse-grained, and coarse-grained with negation. While transformer-based models (e.g., BERT-cased) achieve strong performance on document-level coarse-grained extraction (F1 = 0.943), performance drops notably for fine-grained entity-level tasks (e.g., F1 = 0.675), particularly for rare events and complex attributes. These results demonstrate that despite high-level scores, significant challenges remain in detecting underrepresented AEs and capturing nuanced clinical language. Developed within a Trusted Research Environment (TRE), the dataset is available upon request via DataLoch and serves as a robust benchmark for evaluating AE extraction methods and supporting future cross-dataset generalisation.

cs.CL↗

arXiv2025

Investigating the Capabilities and Limitations of Machine Learning for Identifying Bias in English Language Data with Information and Heritage Professionals

Despite numerous efforts to mitigate their biases, ML systems continue to harm already-marginalized people. While predominant ML approaches assume bias can be removed and fair models can be created, we show that these are not always possible, nor desirable, goals. We reframe the problem of ML bias by creating models to identify biased language, drawing attention to a dataset's biases rather than trying to remove them. Then, through a workshop, we evaluated the models for a specific use case: workflows of information and heritage professionals. Our findings demonstrate the limitations of ML for identifying bias due to its contextual nature, the way in which approaches to mitigating it can simultaneously privilege and oppress different communities, and its inevitability. We demonstrate the need to expand ML approaches to bias and fairness, providing a mixed-methods approach to investigating the feasibility of removing bias or achieving fairness in a given ML use case.

cs.CL↗

arXiv2024

Evaluating and Adapting Large Language Models to Represent Folktales in Low-Resource Languages

Folktales are a rich resource of knowledge about the society and culture of a civilisation. Digital folklore research aims to use automated techniques to better understand these folktales, and it relies on abstract representations of the textual data. Although a number of large language models (LLMs) claim to be able to represent low-resource langauges such as Irish and Gaelic, we present two classification tasks to explore how useful these representations are, and three adaptations to improve the performance of these models. We find that adapting the models to work with longer sequences, and continuing pre-training on the domain of folktales improves classification performance, although these findings are tempered by the impressive performance of a baseline SVM with non-contextual features.

cs.CL↗

arXiv2024

DeCoRe: Decoding by Contrasting Retrieval Heads to Mitigate Hallucinations

Large Language Models (LLMs) often hallucinate, producing unfaithful or factually incorrect outputs by misrepresenting the provided context or incorrectly recalling internal knowledge. Recent studies have identified specific attention heads within the Transformer architecture, known as retrieval heads, responsible for extracting relevant contextual information. We hypothesise that masking these retrieval heads can induce hallucinations and that contrasting the outputs of the base LLM and the masked LLM can reduce hallucinations. To this end, we propose Decoding by Contrasting Retrieval Heads (DeCoRe), a novel training-free decoding strategy that amplifies information found in the context and model parameters. DeCoRe mitigates potentially hallucinated responses by dynamically contrasting the outputs of the base LLM and the masked LLM, using conditional entropy as a guide. Our extensive experiments confirm that DeCoRe significantly improves performance on tasks requiring high contextual faithfulness, such as summarisation (XSum by 18.6%), instruction following (MemoTrap by 10.9%), and open-book question answering (NQ-Open by 2.4% and NQ-Swap by 5.5%).

cs.CL↗

arXiv2024

Automated neuroradiological support systems for multiple cerebrovascular disease markers -- A systematic review and meta-analysis

Cerebrovascular diseases (CVD) can lead to stroke and dementia. Stroke is the second leading cause of death world wide and dementia incidence is increasing by the year. There are several markers of CVD that are visible on brain imaging, including: white matter hyperintensities (WMH), acute and chronic ischaemic stroke lesions (ISL), lacunes, enlarged perivascular spaces (PVS), acute and chronic haemorrhagic lesions, and cerebral microbleeds (CMB). Brain atrophy also occurs in CVD. These markers are important for patient management and intervention, since they indicate elevated risk of future stroke and dementia. We systematically reviewed automated systems designed to support radiologists reporting on these CVD imaging findings. We considered commercially available software and research publications which identify at least two CVD markers. In total, we included 29 commercial products and 13 research publications. Two distinct types of commercial support system were available: those which identify acute stroke lesions (haemorrhagic and ischaemic) from computed tomography (CT) scans, mainly for the purpose of patient triage; and those which measure WMH and atrophy regionally and longitudinally. In research, WMH and ISL were the markers most frequently analysed together, from magnetic resonance imaging (MRI) scans; lacunes and PVS were each targeted only twice and CMB only once. For stroke, commercially available systems largely support the emergency setting, whilst research systems consider also follow-up and routine scans. The systems to quantify WMH and atrophy are focused on neurodegenerative disease support, where these CVD markers are also of significance. There are currently no openly validated systems, commercially, or in research, performing a comprehensive joint analysis of all CVD markers (WMH, ISL, lacunes, PVS, haemorrhagic lesions, CMB, and atrophy).

physics.med-ph↗

arXiv2024

Perceptions of Edinburgh: Capturing Neighbourhood Characteristics by Clustering Geoparsed Local News

The communities that we live in affect our health in ways that are complex and hard to define. Moreover, our understanding of the place-based processes affecting health and inequalities is limited. This undermines the development of robust policy interventions to improve local health and well-being. News media provides social and community information that may be useful in health studies. Here we propose a methodology for characterising neighbourhoods by using local news articles. More specifically, we show how we can use Natural Language Processing (NLP) to unlock further information about neighbourhoods by analysing, geoparsing and clustering news articles. Our work is novel because we combine street-level geoparsing tailored to the locality with clustering of full news articles, enabling a more detailed examination of neighbourhood characteristics. We evaluate our outputs and show via a confluence of evidence, both from a qualitative and a quantitative perspective, that the themes we extract from news articles are sensible and reflect many characteristics of the real world. This is significant because it allows us to better understand the effects of neighbourhoods on health. Our findings on neighbourhood characterisation using news data will support a new generation of place-based research which examines a wider set of spatial processes and how they affect health, enabling new epidemiological research.

cs.IR↗

arXiv2024

A Comparative Study on Patient Language across Therapeutic Domains for Effective Patient Voice Classification in Online Health Discussions

There exists an invisible barrier between healthcare professionals' perception of a patient's clinical experience and the reality. This barrier may be induced by the environment that hinders patients from sharing their experiences openly with healthcare professionals. As patients are observed to discuss and exchange knowledge more candidly on social media, valuable insights can be leveraged from these platforms. However, the abundance of non-patient posts on social media necessitates filtering out such irrelevant content to distinguish the genuine voices of patients, a task we refer to as patient voice classification. In this study, we analyse the importance of linguistic characteristics in accurately classifying patient voices. Our findings underscore the essential role of linguistic and statistical text similarity analysis in identifying common patterns among patient groups. These results allude to even starker differences in the way patients express themselves at a disease level and across various therapeutic domains. Additionally, we fine-tuned a pre-trained Language Model on the combined datasets with similar linguistic patterns, resulting in a highly accurate automatic patient voice classification. Being the pioneering study on the topic, our focus on extracting authentic patient experiences from social media stands as a crucial step towards advancing healthcare standards and fostering a patient-centric approach.

cs.CL↗

arXiv2024

Infusing clinical knowledge into tokenisers for language models

This study introduces a novel knowledge enhanced tokenisation mechanism, K-Tokeniser, for clinical text processing. Technically, at initialisation stage, K-Tokeniser populates global representations of tokens based on semantic types of domain concepts (such as drugs or diseases) from either a domain ontology like Unified Medical Language System or the training data of the task related corpus. At training or inference stage, sentence level localised context will be utilised for choosing the optimal global token representation to realise the semantic-based tokenisation. To avoid pretraining using the new tokeniser, an embedding initialisation approach is proposed to generate representations for new tokens. Using three transformer-based language models, a comprehensive set of experiments are conducted on four real-world datasets for evaluating K-Tokeniser in a wide range of clinical text analytics tasks including clinical concept and relation extraction, automated clinical coding, clinical phenotype identification, and clinical research article classification. Overall, our models demonstrate consistent improvements over their counterparts in all tasks. In particular, substantial improvements are observed in the automated clinical coding task with 13\% increase on Micro $F_1$ score. Furthermore, K-Tokeniser also shows significant capacities in facilitating quicker converge of language models. Specifically, using K-Tokeniser, the language models would only require 50\% of the training data to achieve the best performance of the baseline tokeniser using all training data in the concept extraction task and less than 20\% of the data for the automated coding task. It is worth mentioning that all these improvements require no pre-training process, making the approach generalisable.

cs.CL↗

arXiv2024

Edinburgh Clinical NLP at MEDIQA-CORR 2024: Guiding Large Language Models with Hints

The MEDIQA-CORR 2024 shared task aims to assess the ability of Large Language Models (LLMs) to identify and correct medical errors in clinical notes. In this study, we evaluate the capability of general LLMs, specifically GPT-3.5 and GPT-4, to identify and correct medical errors with multiple prompting strategies. Recognising the limitation of LLMs in generating accurate corrections only via prompting strategies, we propose incorporating error-span predictions from a smaller, fine-tuned model in two ways: 1) by presenting it as a hint in the prompt and 2) by framing it as multiple-choice questions from which the LLM can choose the best correction. We found that our proposed prompting strategies significantly improve the LLM's ability to generate corrections. Our best-performing solution with 8-shot + CoT + hints ranked sixth in the shared task leaderboard. Additionally, our comprehensive analyses show the impact of the location of the error sentence, the prompted role, and the position of the multiple-choice option on the accuracy of the LLM. This prompts further questions about the readiness of LLM to be implemented in real-world clinical settings.

cs.CL↗

arXiv2024

Edinburgh Clinical NLP at SemEval-2024 Task 2: Fine-tune your model unless you have access to GPT-4

The NLI4CT task assesses Natural Language Inference systems in predicting whether hypotheses entail or contradict evidence from Clinical Trial Reports. In this study, we evaluate various Large Language Models (LLMs) with multiple strategies, including Chain-of-Thought, In-Context Learning, and Parameter-Efficient Fine-Tuning (PEFT). We propose a PEFT method to improve the consistency of LLMs by merging adapters that were fine-tuned separately using triplet and language modelling objectives. We found that merging the two PEFT adapters improves the F1 score (+0.0346) and consistency (+0.152) of the LLMs. However, our novel methods did not produce more accurate results than GPT-4 in terms of faithfulness and consistency. Averaging the three metrics, GPT-4 ranks joint-first in the competition with 0.8328. Finally, our contamination analysis with GPT-4 indicates that there was no test data leakage.

cs.CL↗

arXiv2024

Can GPT-3.5 Generate and Code Discharge Summaries?

Objective: To investigate GPT-3.5 in generating and coding medical documents with ICD-10 codes for data augmentation on low-resources labels. Materials and Methods: Employing GPT-3.5 we generated and coded 9,606 discharge summaries based on lists of ICD-10 code descriptions of patients with infrequent (generation) codes within the MIMIC-IV dataset. Combined with the baseline training set, this formed an augmented training set. Neural coding models were trained on baseline and augmented data and evaluated on a MIMIC-IV test set. We report micro- and macro-F1 scores on the full codeset, generation codes, and their families. Weak Hierarchical Confusion Matrices were employed to determine within-family and outside-of-family coding errors in the latter codesets. The coding performance of GPT-3.5 was evaluated both on prompt-guided self-generated data and real MIMIC-IV data. Clinical professionals evaluated the clinical acceptability of the generated documents. Results: Augmentation slightly hinders the overall performance of the models but improves performance for the generation candidate codes and their families, including one unseen in the baseline training data. Augmented models display lower out-of-family error rates. GPT-3.5 can identify ICD-10 codes by the prompted descriptions, but performs poorly on real data. Evaluators note the correctness of generated concepts while suffering in variety, supporting information, and narrative. Discussion and Conclusion: GPT-3.5 alone is unsuitable for ICD-10 coding. Augmentation positively affects generation code families but mainly benefits codes with existing examples. Augmentation reduces out-of-family errors. Discharge summaries generated by GPT-3.5 state prompted concepts correctly but lack variety, and authenticity in narratives. They are unsuitable for clinical practice.

cs.CL↗

arXiv2023

Classifying patient voice in social media data using neural networks: A comparison of AI models on different data sources and therapeutic domains

It is essential that healthcare professionals and members of the healthcare community can access and easily understand patient experiences in the real world, so that care standards can be improved and driven towards personalised drug treatment. Social media platforms and message boards are deemed suitable sources of patient experience information, as patients have been observed to discuss and exchange knowledge, look for and provide support online. This paper tests the hypothesis that not all online patient experience information can be treated and collected in the same way, as a result of the inherent differences in the way individuals talk about their journeys, in different therapeutic domains and or data sources. We used linguistic analysis to understand and identify similarities between datasets, across patient language, between data sources (Reddit, SocialGist) and therapeutic domains (cardiovascular, oncology, immunology, neurology). We detected common vocabulary used by patients in the same therapeutic domain across data sources, except for immunology patients, who use unique vocabulary between the two data sources, and compared to all other datasets. We combined linguistically similar datasets to train classifiers (CNN, transformer) to accurately identify patient experience posts from social media, a task we refer to as patient voice classification. The cardiovascular and neurology transformer classifiers perform the best in their respective comparisons for the Reddit data source, achieving F1-scores of 0.865 and 1.0 respectively. The overall best performing classifier is the transformer classifier trained on all data collected for this experiment, achieving F1-scores ranging between 0.863 and 0.995 across all therapeutic domain and data source specific test datasets.

cs.CL↗

arXiv2023

Parameter-Efficient Fine-Tuning of LLaMA for the Clinical Domain

Adapting pretrained language models to novel domains, such as clinical applications, traditionally involves retraining their entire set of parameters. Parameter-Efficient Fine-Tuning (PEFT) techniques for fine-tuning language models significantly reduce computational requirements by selectively fine-tuning small subsets of parameters. In this study, we propose a two-step PEFT framework and evaluate it in the clinical domain. Our approach combines a specialised PEFT adapter layer designed for clinical domain adaptation with another adapter specialised for downstream tasks. We evaluate the framework on multiple clinical outcome prediction datasets, comparing it to clinically trained language models. Our framework achieves a better AUROC score averaged across all clinical downstream tasks compared to clinical language models. In particular, we observe large improvements of 4-5% AUROC in large-scale multilabel classification tasks, such as diagnoses and procedures classification. To our knowledge, this study is the first to provide an extensive empirical analysis of the interplay between PEFT techniques and domain adaptation in an important real-world domain of clinical applications.

cs.CL↗

arXiv2022

Ontology-Driven and Weakly Supervised Rare Disease Identification from Clinical Notes

Computational text phenotyping is the practice of identifying patients with certain disorders and traits from clinical notes. Rare diseases are challenging to be identified due to few cases available for machine learning and the need for data annotation from domain experts. We propose a method using ontologies and weak supervision, with recent pre-trained contextual representations from Bi-directional Transformers (e.g. BERT). The ontology-based framework includes two steps: (i) Text-to-UMLS, extracting phenotypes by contextually linking mentions to concepts in Unified Medical Language System (UMLS), with a Named Entity Recognition and Linking (NER+L) tool, SemEHR, and weak supervision with customised rules and contextual mention representation; (ii) UMLS-to-ORDO, matching UMLS concepts to rare diseases in Orphanet Rare Disease Ontology (ORDO). The weakly supervised approach is proposed to learn a phenotype confirmation model to improve Text-to-UMLS linking, without annotated data from domain experts. We evaluated the approach on three clinical datasets, MIMIC-III discharge summaries, MIMIC-III radiology reports, and NHS Tayside brain imaging reports from two institutions in the US and the UK, with annotations. The improvements in the precision were pronounced (by over 30% to 50% absolute score for Text-to-UMLS linking), with almost no loss of recall compared to the existing NER+L tool, SemEHR. Results on radiology reports from MIMIC-III and NHS Tayside were consistent with the discharge summaries. The overall pipeline processing clinical notes can extract rare disease cases, mostly uncaptured in structured data (manually assigned ICD codes). We discuss the usefulness of the weak supervision approach and propose directions for future studies.

cs.CL↗