Retrieval-augmented generation (RAG) can substantially enhance the performance of LLMs on knowledge-intensive tasks. Various RAG paradigms—including vanilla, planning-based, and iterative RAG—all depend on a robust retriever, yet existing retrievers rely heavily on public knowledge and often falter when faced with domain-specific queries. To address these limitations, we introduce DRAGON, a framework that combines a data-construction modeling approach with a scalable synthetic data-generation pipeline, specifically designed to optimize domain-specific retrieval performance and bolster retriever robustness. To evaluate RAG performance on domain-specific RAGs, we propose DRAGONBench, a benchmark spanning 8 domain-specific document collections across 4 distinct fields and featuring a wide spectrum of query complexities, answerability, and hops. Leveraging DRAGON, we generate a large-scale synthetic dataset—encompassing both single-hop and multi-hop queries—to enrich retriever training. Extensive experiments demonstrate that retrievers trained on this data yield significant performance gains and exhibit strong cross-domain generalization. Moreover, when our optimized retrievers are integrated into vanilla, planning-based, and iterative RAG paradigms, we observe consistent end-to-end improvements in system accuracy.

Modern language models excel at factual reasoning but struggle with value diversity: the multiplicity of plausible human perspectives. Tasks such as hate speech or sexism detection expose this limitation, where human disagreement captures the diversity of perspectives that models need to account for, rather than dataset noise. In this paper, we explore whether multi-perspective in-context learning (ICL) can align large language models (LLMs) with this diversity without parameter updates. We evaluate four LLMs on five datasets across three languages (English, Arabic, Italian), considering three label-space representations (aggregated hard, disaggregated hard, and disaggregated soft) and five demonstration selection and ordering strategies. Our multi-perspective approach outperforms standard prompting on aggregated English labels, while disaggregated soft predictions better align with human judgments in Arabic and Italian datasets.These findings highlight the importance of perspective-aware LLMs for reducing bias and polarization, while also revealing the challenges of applying ICL to socially sensitive tasks. We further probe the model faithfulness using eXplainable AI (XAI), offering insights into how LLMs handle human disagreement.

Extremely low-resource languages, especially those written in rare scripts, remain largely unsupported by large language models (LLMs). This is due in part to compounding factors such as the lack of training data. This paper delivers the first comprehensive analysis of whether LLMs can acquire such languages purely via in-context learning (ICL), with or without auxiliary alignment signals, and how these methods compare to parameter-efficient fine-tuning (PEFT). We systematically evaluate 20 under-represented languages across three state-of-the-art multilingual LLMs. Our findings highlight the limitation of PEFT when both language and its script are extremely under-represented by the LLM. In contrast, zero-shot ICL with language alignment is impressively effective on extremely low-resource languages, while few-shot ICL or PEFT is more beneficial for languages relatively better represented by LLMs. For LLM practitioners working on extremely low-resource languages, we summarise guidelines grounded by our results on adapting LLMs to low-resource languages, e.g., avoiding fine-tuning a multilingual model on languages of unseen scripts.

In-context learning (ICL) performance is highly sensitive to prompt design, yet the impact of class label options (e.g. lexicon or order) in zero-shot classification remains underexplored. This study proposes LOADS (Label set Optimization via Activation Distribution kurtosiS), a post-hoc method for selecting optimal label sets in zero-shot ICL with large language models (LLMs).LOADS is built upon the observations in our empirical analysis, the first to systematically examine how label option design (i.e., lexical choice, order, and elaboration) impacts classification performance. This analysis shows that the lexical choice of the labels in the prompt (such as agree vs. support in stance classification) plays an important role in both model performance and model’s sensitivity to the label order. A further investigation demonstrates that optimal label words tend to activate fewer outlier neurons in LLMs’ feed-forward networks. LOADS then leverages kurtosis to measure the neuron activation distribution for label selection, requiring only a single forward pass without gradient propagation or labelled data. The LOADS-selected label words consistently demonstrate effectiveness for zero-shot ICL across classification tasks, datasets, models and languages, achieving maximum performance gain from 0.54 to 0.76 compared to the conventional approach of using original dataset label words.

Considering a conversation thread, rumour stance classification aims to identify the opinion (e.g. agree or disagree) of replies towards a target (rumour story). Although the target is expected to be an essential component in traditional stance classification, we show that rumour stance classification datasets contain a considerable amount of real-world data whose stance could be naturally inferred directly from the replies, contributing to the strong performance of the supervised models without awareness of the target. We find that current target-aware models underperform in cases where the context of the target is crucial. Finally, we propose a simple yet effective framework to enhance reasoning with the targets, achieving state-of-the-art performance on two benchmark datasets.

Vaccine hesitancy is widespread, despite the government’s information campaigns and the efforts of the World Health Organisation (WHO). Categorising the topics within vaccine-related narratives is crucial to understand the concerns expressed in discussions and identify the specific issues that contribute to vaccine hesitancy. This paper addresses the need for monitoring and analysing vaccine narratives online by introducing a novel vaccine narrative classification task, which categorises COVID-19 vaccine claims into one of seven categories. Following a data augmentation approach, we first construct a novel dataset for this new classification task, focusing on the minority classes. We also make use of fact-checker annotated data. The paper also presents a neural vaccine narrative classifier that achieves an accuracy of 84% under cross-validation. The classifier is publicly available for researchers and journalists.

The use of visually-rich documents in various fields has created a demand for Document AI models that can read and comprehend documents like humans, which requires the overcoming of technical, linguistic, and cognitive barriers. Unfortunately, the lack of appropriate datasets has significantly hindered advancements in the field. To address this issue, we introduce DocTrack, a visually-rich document dataset really aligned with human eye-movement information using eye-tracking technology. This dataset can be used to investigate the challenges mentioned above. Additionally, we explore the impact of human reading order on document understanding tasks and examine what would happen if a machine reads in the same order as a human. Our results suggest that although Document AI models have made significant progresses, they still have a long way to go before they can read visually richer documents as accurately, continuously, and flexibly as humans do. These findings have potential implications for future research and development of document intelligence.

In this paper, we address the limitations of the common data annotation and training methods for objective single-label classification tasks. Typically, when annotating such tasks annotators are only asked to provide a single label for each sample and annotator disagreement is discarded when a final hard label is decided through majority voting. We challenge this traditional approach, acknowledging that determining the appropriate label can be difficult due to the ambiguity and lack of context in the data samples. Rather than discarding the information from such ambiguous annotations, our soft label method makes use of them for training. Our findings indicate that additional annotator information, such as confidence, secondary label and disagreement, can be used to effectively generate soft labels. Training classifiers with these soft labels then leads to improved performance and calibration on the hard label test set.

This paper describes the second-placed system on the leaderboard of SemEval-2022 Task 8: Multilingual News Article Similarity. We propose an entity-enriched Siamese Transformer which computes news article similarity based on different sub-dimensions, such as the shared narrative, entities, location and time of the event discussed in the news article. Our system exploits a Siamese network architecture using a Transformer encoder to learn document-level representations for the purpose of capturing the narrative together with the auxiliary entity-based features extracted from the news articles. The intuition behind using all these features together is to capture the similarity between news articles at different granularity levels and to assess the extent to which different news outlets write about “the same events”. Our experimental results and detailed ablation study demonstrate the effectiveness and the validity of our proposed method.

Meta learning aims to optimize the model’s capability to generalize to new tasks and domains. Lacking a data-efficient way to create meta training tasks has prevented the application of meta-learning to the real-world few shot learning scenarios. Recent studies have proposed unsupervised approaches to create meta-training tasks from unlabeled data for free, e.g., the SMLMT method (Bansal et al., 2020a) constructs unsupervised multi-class classification tasks from the unlabeled text by randomly masking words in the sentence and let the meta learner choose which word to fill in the blank. This study proposes a semi-supervised meta-learning approach that incorporates both the representation power of large pre-trained language models and the generalization capability of prototypical networks enhanced by SMLMT. The semi-supervised meta training approach avoids overfitting prototypical networks on a small number of labeled training examples and quickly learns cross-domain task-specific representation only from a few supporting examples. By incorporating SMLMT with prototypical networks, the meta learner generalizes better to unseen domains and gains higher accuracy on out-of-scope examples without the heavy lifting of pre-training. We observe significant improvement in few-shot generalization after training only a few epochs on the intent classification tasks evaluated in a multi-domain setting.

Correctly classifying stances of replies can be significantly helpful for the automatic detection and classification of online rumours. One major challenge is that there are considerably more non-relevant replies (comments) than informative ones (supports and denies), making the task highly imbalanced. In this paper we revisit the task of rumour stance classification, aiming to improve the performance over the informative minority classes. We experiment with traditional methods for imbalanced data treatment with feature- and BERT-based classifiers. Our models outperform all systems in RumourEval 2017 shared task and rank second in RumourEval 2019.