Pierre Zweigenbaum

Hui-Syuan Yeh, Thomas Lavergne, Pierre Zweigenbaum

Relation extraction is a core problem for natural language processing in the biomedical domain. Recent research on relation extraction showed that prompt-based learning improves the performance on both fine-tuning on full training set and few-shot training. However, less effort has been made on domain-specific tasks where good prompt design can be even harder. In this paper, we investigate prompting for biomedical relation extraction, with experiments on the ChemProt dataset. We present a simple yet effective method to systematically generate comprehensive prompts that reformulate the relation extraction task as a cloze-test task under a simple prompt formulation. In particular, we experiment with different ranking scores for prompt selection. With BioMed-RoBERTa-base, our results show that prompting-based fine-tuning obtains gains by 14.21 F1 over its regular fine-tuning baseline, and 1.14 F1 over SciFive-Large, the current state-of-the-art on ChemProt. Besides, we find prompt-based learning requires fewer training examples to make reasonable predictions. The results demonstrate the potential of our methods in such a domain-specific relation extraction task.

Lisa Raithel, Philippe Thomas, Roland Roller et al.

In this work, we present the first corpus for German Adverse Drug Reaction (ADR) detection in patient-generated content. The data consists of 4,169 binary annotated documents from a German patient forum, where users talk about health issues and get advice from medical doctors. As is common in social media data in this domain, the class labels of the corpus are very imbalanced. This and a high topic imbalance make it a very challenging dataset, since often, the same symptom can have several causes and is not always related to a medication intake. We aim to encourage further multi-lingual efforts in the domain of ADR detection and provide preliminary experiments for binary classification using different methods of zero- and few-shot learning based on a multi-lingual model. When fine-tuning XLM-RoBERTa first on English patient forum data and then on the new German data, we achieve an F1-score of 37.52 for the positive class. We make the dataset and models publicly available for the community.

Xavier Tannier, Salam Abbara, Rémi Flicoteaux et al.

The development of clinical natural language processing (NLP) systems is severely hampered by the sensitive nature of medical records, which restricts data sharing under stringent privacy regulations, particularly in France and the broader European Union. To address this gap, we introduce PARHAF, a large open-source corpus of clinical documents in French. PARHAF comprises expert-authored clinical reports describing realistic yet entirely fictitious patient cases, making it anonymous and freely shareable by design. The corpus was developed using a structured protocol that combined clinician expertise with epidemiological guidance from the French National Health Data System (SNDS), ensuring broad clinical coverage. A total of 104 medical residents across 18 specialties authored and peer-reviewed the reports following predefined clinical scenarios and document templates. The corpus contains 7394 clinical reports covering 5009 patient cases across a wide range of medical and surgical specialties. It includes a general-purpose component designed to approximate real-world hospitalization distributions, and four specialized subsets that support information-extraction use cases in oncology, infectious diseases, and diagnostic coding. Documents are released under a CC-BY open license, with a portion temporarily embargoed to enable future benchmarking under controlled conditions. PARHAF provides a valuable resource for training and evaluating French clinical language models in a fully privacy-preserving setting, and establishes a replicable methodology for building shareable synthetic clinical corpora in other languages and health systems.

Oumaima El Khettari, Virgile Barthet, Guillaume Hocquet et al.

Accurate short-term mortality prediction in heart failure (HF) remains challenging, particularly when relying on structured electronic health record (EHR) data alone. We evaluate transformer-based models on a French HF cohort, comparing text-only, structured-only, multimodal, and LLM-based approaches. Our results show that enriching clinical text with entity-level representations improves prediction over CLS embeddings alone, and that supervised multimodal fusion of text and structured variables achieves the best overall performance. In contrast, large language models perform inconsistently across modalities and decoding strategies, with text-only prompts outperforming structured or multimodal inputs. These findings highlight that entity-aware multimodal transformers offer the most reliable solution for short-term HF outcome prediction, while current LLM prompting remains limited for clinical decision support.

Mathilde Aguiar, Pierre Zweigenbaum, Nona Naderi

Recruiting patients to participate in clinical trials can be challenging and time-consuming. Usually, participation in a clinical trial is initiated by a healthcare professional and proposed to the patient. Promoting clinical trials directly to patients via online recruitment might help to reach them more efficiently. In this study, we address the case where a patient is initiating their own recruitment process and wants to determine whether they are eligible for a given clinical trial, using their own language to describe their medical profile. To study whether this creates difficulties in the patient trial matching process, we design a new dataset and task, Natural Language Inference for Patient Recruitment (NLI4PR), in which patient language profiles must be matched to clinical trials. We create it by adapting the TREC 2022 Clinical Trial Track dataset, which provides patients' medical profiles, and rephrasing them manually using patient language. We also use the associated clinical trial reports where the patients are either eligible or excluded. We prompt several open-source Large Language Models on our task and achieve from 56.5 to 71.8 of F1 score using patient language, against 64.7 to 73.1 for the same task using medical language. When using patient language, we observe only a small loss in performance for the best model, suggesting that having the patient as a starting point could be adopted to help recruit patients for clinical trials. The corpus and code bases are all freely available on our Github and HuggingFace repositories.

Anfu Tang, Louise Deléger, Robert Bossy et al.

Recently many studies have been conducted on the topic of relation extraction. The DrugProt track at BioCreative VII provides a manually-annotated corpus for the purpose of the development and evaluation of relation extraction systems, in which interactions between chemicals and genes are studied. We describe the ensemble system that we used for our submission, which combines predictions of fine-tuned bioBERT, sciBERT and const-bioBERT models by majority voting. We specifically tested the contribution of syntactic information to relation extraction with BERT. We observed that adding constituentbased syntactic information to BERT improved precision, but decreased recall, since relations rarely seen in the train set were less likely to be predicted by BERT models in which the syntactic information is infused. Our code is available online [https://github.com/Maple177/drugprot-relation-extraction].

Ayoub Hammal, Pierre Zweigenbaum, Caio Corro

Recent works proposed test-time alignment methods that rely on a small aligned model as a proxy that guides the generation of a larger base (unaligned) model. The implicit reward approach skews the large model distribution, whereas the nudging approach defers the generation of the next token to the small aligned model when the large base one is unconfident about its outcome. In this work, we first show that both approaches can be reduced to sampling from similar graphical models, where they differ only in the definition of a rejection criterion (or distribution). Moreover, we argue that the confidence criterion is ill-motivated due to linguistic phenomena like ambiguous phrasing. We propose a novel rejection criterion based on a conservative confidence bet. Experimentally, our novel approach outperforms previous work on several datasets.

Ayoub Hammal, Pierre Zweigenbaum, Caio Corro

Large language models (LLMs) are powerful tools that have found applications beyond human-machine interfaces and chatbots. In particular, their ability to generate reasoning traces motivated their use in many prediction tasks like math question answering. Unfortunately, extracting the final answer in an LLM free-form output is difficult, as it is an information extraction problem on its own. In this work, we introduce suffix-constrained generation, that aims to produce well-formed LLM responses in which final answers follow strict templates and are guaranteed to be trivially parseable. To this end, we introduce several algorithms that are based on greedy search procedures. We experiment on several datasets, and show that our approach allows to guarantee trivial deterministic extraction of the final answer from an LLM output without having a negative impact on results, and even improving them.

Ayoub Hammal, Pierre Zweigenbaum, Caio Corro

Several previous works concluded that the largest part of generation capabilities of large language models (LLM) are learned (early) during pre-training. However, LLMs still require further alignment to adhere to downstream task requirements and stylistic preferences, among other desired properties. As LLMs continue to scale in terms of size, the computational cost of alignment procedures increase prohibitively. In this work, we propose a novel approach to circumvent these costs via proxy-based test-time alignment, i.e. using guidance from a small aligned model. Our approach can be described as token-specific cascading method, where the token-specific deferral rule is reduced to 0-1 knapsack problem. In this setting, we derive primal and dual approximations of the optimal deferral decision. We experimentally show the benefits of our method both in task performance and speculative decoding speed.

Lisa Raithel, Hui-Syuan Yeh, Shuntaro Yada et al.

User-generated data sources have gained significance in uncovering Adverse Drug Reactions (ADRs), with an increasing number of discussions occurring in the digital world. However, the existing clinical corpora predominantly revolve around scientific articles in English. This work presents a multilingual corpus of texts concerning ADRs gathered from diverse sources, including patient fora, social media, and clinical reports in German, French, and Japanese. Our corpus contains annotations covering 12 entity types, four attribute types, and 13 relation types. It contributes to the development of real-world multilingual language models for healthcare. We provide statistics to highlight certain challenges associated with the corpus and conduct preliminary experiments resulting in strong baselines for extracting entities and relations between these entities, both within and across languages.

Mathilde Aguiar, Pierre Zweigenbaum, Nona Naderi

This paper describes our submission to Task 2 of SemEval-2024: Safe Biomedical Natural Language Inference for Clinical Trials. The Multi-evidence Natural Language Inference for Clinical Trial Data (NLI4CT) consists of a Textual Entailment (TE) task focused on the evaluation of the consistency and faithfulness of Natural Language Inference (NLI) models applied to Clinical Trial Reports (CTR). We test 2 distinct approaches, one based on finetuning and ensembling Masked Language Models and the other based on prompting Large Language Models using templates, in particular, using Chain-Of-Thought and Contrastive Chain-Of-Thought. Prompting Flan-T5-large in a 2-shot setting leads to our best system that achieves 0.57 F1 score, 0.64 Faithfulness, and 0.56 Consistency.

Anfu Tang, Claire Nédellec, Pierre Zweigenbaum et al.

We investigate a method to extract relations from texts based on global alignment and syntactic information. Combined with SVM, this method is shown to have a performance comparable or even better than LSTM on two RE tasks.

Hicham El Boukkouri, Olivier Ferret, Thomas Lavergne et al.

Due to the compelling improvements brought by BERT, many recent representation models adopted the Transformer architecture as their main building block, consequently inheriting the wordpiece tokenization system despite it not being intrinsically linked to the notion of Transformers. While this system is thought to achieve a good balance between the flexibility of characters and the efficiency of full words, using predefined wordpiece vocabularies from the general domain is not always suitable, especially when building models for specialized domains (e.g., the medical domain). Moreover, adopting a wordpiece tokenization shifts the focus from the word level to the subword level, making the models conceptually more complex and arguably less convenient in practice. For these reasons, we propose CharacterBERT, a new variant of BERT that drops the wordpiece system altogether and uses a Character-CNN module instead to represent entire words by consulting their characters. We show that this new model improves the performance of BERT on a variety of medical domain tasks while at the same time producing robust, word-level and open-vocabulary representations.

Zheng Zhang, Ruiqing Yin, Jun Zhu et al.

Recent work in cross-lingual contextual word embedding learning cannot handle multi-sense words well. In this work, we explore the characteristics of contextual word embeddings and show the link between contextual word embeddings and word senses. We propose two improving solutions by considering contextual multi-sense word embeddings as noise (removal) and by generating cluster level average anchor embeddings for contextual multi-sense word embeddings (replacement). Experiments show that our solutions can improve the supervised contextual word embeddings alignment for multi-sense words in a microscopic perspective without hurting the macroscopic performance on the bilingual lexicon induction task. For unsupervised alignment, our methods significantly improve the performance on the bilingual lexicon induction task for more than 10 points.