We describe team blue’s participation across six SMM4H-HeaRD 2026 shared tasks spanning multilingual adverse drug event detection (Task 1), influenza vaccine effectiveness estimation (Task 3), patient metadata classification (Task 5), TNM cancer staging (Task 6), opioid impact span detection (Task 7), and multilingual clinical NER with cross-lingual annotation projection (Task 8). Despite the heterogeneity of these tasks, binary, multi-class, multi-label, and sequence-labelling, our systems share three recurring design principles: (i) inverse-frequency class weighting to handle severe imbalance, (ii) multi-seed and/or multi-backbone ensembling to reduce variance, and (iii) post-hoc calibration of decision boundaries. Key results include micro-F1 of 0.990 on TNM staging (Task 6), 0.872/0.918 on flu vaccination/test classification surpassing the 70B CoT baseline on vaccination (Task 3), F1 of 0.764 on patient metadata approaching the fine-tuning benchmark of 0.776 (Task 5), and competitive performance on ADE detection (Task 1, F1 = 0.580), opioid spans (Task 7, relaxed F1 = 0.59), and multilingual clinical NER (Task 8, strict F1 0.20–0.41 across 7 languages).

This paper outlines the method submitted by team blue for the SemEval-2026 Task 5: Rating Plausibility of Word Senses in Ambiguous Sentences through Narrative (AmbiStory). The task requires predicting reasonable scores that match human thoughts and judgments instead of just picking a single correct sense as the output. This means that contextual reasoning with fine-grain contextual modeling is vital. In order to tackle this problem, we suggest a BERT-based cross-encoder regression model. This model encodes the entire narrative context, which includes the precontext, the ambiguous sentence, and the ending, along with candidate sense definitions and example usages. Unlike bi-encoder sentence-level methods, our model allows for token-level interaction between story cues and sense meanings. This interaction helps capture subtle narrative disambiguation signals. We conduct a systematic exploration of model architectures and training strategies, progressing from a sentence-transformer baseline to an optimised BERT cross-encoder. On the development set, our best configuration achieves a Spearman rank correlation of 0.66. On the official test set, the system achieves a Spearman correlation of 0.4866 and an Accuracy-within-Standard-Deviation of 0.6613, substantially outperforming sentence-transformer bi-encoder baselines.

This paper describes the system submitted by team blue for SemEval-2026 Task 4: Narrative Story Similarity and Narrative Representation Learning, with a primary focus on the Pairwise Similarity subtask (Track A). The core challenge of this task lies in identifying deep structural alignments between stories, which is fundamentally hindered by the restricted context windows of standard transformer architecturesthat truncate narratives before reaching critical plot resolutions. To overcome this context bottleneck, we propose a hybrid ensemble architecture designed to capture extended narrative arcs. Our approach synergizes a cross-encoder (Jina Reranker v2), which processes long inputs via a sliding-window strategy over 1,024-token chunks, to evaluate the global "course of action," with a semantic bi-encoder (RoBERTa-Large) to validate local tonal consistency. This dual-stream system achieved a Pearson correlation score of 0.63, demonstrating that processing narrative content beyond the 512-token truncation boundary is strictly necessary for accurate pairwise narrative comparison.