Retrieval Augmented Generation (RAG) has advanced Question Answering (QA) by connecting Large Language Models (LLMs) to external knowledge. However, these systems can still produce answers that are unsupported, lack clear traceability, or misattribute information — a critical issue in the biomedical domain where accuracy, trust and control are essential. We introduce TripleCheck, a post-hoc framework that breaks down an LLM’s answer into factual triples and checks each against both the retrieved context and a biomedical knowledge graph. By highlighting which statements are supported, traceable, or correctly attributed, TripleCheck enables users to spot gaps, unsupported claims, and misattributions, prompting more careful follow up. We present the TripleCheck framework, evaluate it on the SciFact benchmark, analyze its limitations, and share preliminary expert feedback. Results show that TripleCheck provides nuanced insight, potentially supporting greater trust and safer AI adoption in biomedical applications.

The one-sided focus on English in previous studies of gender bias in NLP misses out on opportunities in other languages: English challenge datasets such as GAP and WinoGender highlight model preferences that are “hallucinatory”, e.g., disambiguating gender-ambiguous occurrences of ‘doctor’ as male doctors. We show that for languages with type B reflexivization, e.g., Swedish and Russian, we can construct multi-task challenge datasets for detecting gender bias that lead to unambiguously wrong model predictions: In these languages, the direct translation of ‘the doctor removed his mask’ is not ambiguous between a coreferential reading and a disjoint reading. Instead, the coreferential reading requires a non-gendered pronoun, and the gendered, possessive pronouns are anti-reflexive. We present a multilingual, multi-task challenge dataset, which spans four languages and four NLP tasks and focuses only on this phenomenon. We find evidence for gender bias across all task-language combinations and correlate model bias with national labor market statistics.

Unresolved coreference is a bottleneck for relation extraction, and high-quality coreference resolvers may produce an output that makes it a lot easier to extract knowledge triples. We show how to improve coreference resolvers by forwarding their input to a relation extraction system and reward the resolvers for producing triples that are found in knowledge bases. Since relation extraction systems can rely on different forms of supervision and be biased in different ways, we obtain the best performance, improving over the state of the art, using multi-task reinforcement learning.

Neural machine translation models have little inductive bias, which can be a disadvantage in low-resource scenarios. Neural models have to be trained on large amounts of data and have been shown to perform poorly when only limited data is available. We show that using naive regularization methods, based on sentence length, punctuation and word frequencies, to penalize translations that are very different from the input sentences, consistently improves the translation quality across multiple low-resource languages. We experiment with 12 language pairs, varying the training data size between 17k to 230k sentence pairs. Our best regularizer achieves an average increase of 1.5 BLEU score and 1.0 TER score across all the language pairs. For example, we achieve a BLEU score of 26.70 on the IWSLT15 English–Vietnamese translation task simply by using relative differences in punctuation as a regularizer.

This work describes the system presented by the CoAStaL Natural Language Processing group at University of Copenhagen. The main system we present uses the same attention mechanism presented in (Yang et al., 2016). Our overall model architecture is also inspired by their hierarchical classification model and adapted to deal with classification in dialogue by encoding information at the turn level. We use different encodings for each turn to create a more expressive representation of dialogue context which is then fed into our classifier. We also define a custom preprocessing step in order to deal with language commonly used in interactions across many social media outlets. Our proposed system achieves a micro F1 score of 0.7340 on the test set and shows significant gains in performance compared to a system using dialogue level encoding.