We present our participation in the SOMD 2026 shared task on cross-document software mention coreference resolution, where our systems ranked second across all three subtasks. We compare two fine-tuning-free approaches: Fuzzy Matching (FM), a lexical string-similarity method, and Context Aware Representations (CAR), which combines mention-level and document-level embeddings. Both achieve competitive performance across all subtasks (CoNLL F1 of 0.94–0.96), with CAR consistently outperforming FM by 1 point on the official test set, consistent with the high surface regularity of software names, which reduces the need for complex semantic reasoning. A controlled noise-injection study reveals complementary failure modes: as boundary noise increases, CAR loses only 0.07 F1 points from clean to fully corrupted input, compared to 0.20 for FM, whereas under mention substitution, FM degrades more gracefully (0.52 vs. 0.63). Our inference-time analysis shows that FM scales superlinearly with corpus size, whereas CAR scales approximately linearly, making CAR the more efficient choice at large scale. These findings suggest that system selection should be informed by both the noise profile of the upstream mention detector and the scale of the target corpus. We release our code to support future work on this underexplored task.

Scientific multi-label text classification suffers from extreme class imbalance, where specialized terminology exhibits severe power-law distributions that challenge standard classification approaches. Existing scientific corpora lack comprehensive controlled vocabularies, focusing instead on broad categories and limiting systematic study of extreme imbalance. We introduce AstroConcepts, a corpus of English abstracts from 21,702 published astrophysics papers, labeled with 2,367 concepts from the Unified Astronomy Thesaurus. The corpus exhibits severe label imbalance, with 76 % of concepts having fewer than 50 training examples. By releasing this resource, we enable systematic study of extreme class imbalance in scientific domains and establish strong baselines across traditional, neural, and vocabulary-constrained LLM methods. Our evaluation reveals three key patterns that provide new insights into scientific text classification. First, vocabulary-constrained LLMs achieve competitive performance relative to domain-adapted models in astrophysics classification, suggesting a potential for parameter-efficient approaches. Second, domain adaptation yields relatively larger improvements for rare, specialized terminology, although absolute performance remains limited across all methods. Third, we propose frequency-stratified evaluation to reveal performance patterns that are hidden by aggregate scores, thereby making robustness assessment central to scientific multi-label evaluation. These results offer actionable insights for scientific NLP and establish benchmarks for research on extreme imbalance.

To evaluate the scientific influence of observational facilities, astronomers examine the body of publications that have utilized data from those facilities. This depends on curated bibliographies that annotate and connect data products to the corresponding literature, enabling bibliometric analyses to quantify data impact. Compiling such bibliographies is a demanding process that requires expert curators to scan the literature for relevant names, acronyms, and identifiers, and then to determine whether and how specific observations contributed to each publication. These bibliographies have value beyond impact assessment: for research scientists, explicit links between data and literature form an essential pathway for discovering and accessing data. Accordingly, by building on the work of librarians and archivists, telescope bibliographies can be repurposed to directly support scientific inquiry. In this context, we present the Telescope Reference and Astronomy Categorization Shared task (TRACS) and its accompanying dataset, which comprises more than 89,000 publicly available English-language texts drawn from space telescope bibliographies. These texts are labeled according to a new, compact taxonomy developed in consultation with experienced bibliographers.

The Workshop for Artificial Intelligence for Scientific Publications (WASP), formerly Workshop on Information Extraction from Scientific Publications (WIESP), started in 2022 to provide a platform for researchers to discuss research on information extraction, mining, generation, and knowledge discovery from scientific publications using Natural Language Processing and Machine Learning techniques. The third WASP workshop was held at the 14th International Joint Conference on Natural Language Processing and 4th Asia-Pacific Chapter of the Association for Computational Linguistics in Mumbai, India on December 23rd, 2025, as a hybrid event. The WASP workshop saw great interest, with 29 submissions, of which 16 were accepted. The program consisted of the contributed research talks, 2 keynote talks, a panel discussion, and one shared task, Telescope Reference and Astronomy Categorization Shared task (TRACS).

Large language models (LLMs) trained on general domain corpora showed remarkable results on natural language processing (NLP) tasks. However, previous research demonstrated LLMs trained using domain-focused corpora perform better on specialized tasks. Inspired by this insight, we developed INDUS, a comprehensive suite of LLMs tailored for the closely-related domains of Earth science, biology, physics, heliophysics, planetary sciences and astrophysics, and trained using curated scientific corpora drawn from diverse data sources. The suite of models include: (1) an encoder model trained using domain-specific vocabulary and corpora to address NLP tasks, (2) a contrastive-learning based text embedding model trained using a diverse set of datasets to address information retrieval tasks and (3) smaller versions of these models created using knowledge distillation for applications which have latency or resource constraints. We also created three new scientific benchmark datasets, Climate-Change NER (entity-recognition), NASA-QA (extractive QA) and NASA-IR (IR) to accelerate research in these multi-disciplinary fields. We show that our models outperform both general-purpose (RoBERTa) and domain- specific (SciBERT) encoders on these new tasks as well as existing tasks in the domains of interest. Furthermore, we demonstrate the use of these models in two industrial settings- as a retrieval model for large-scale vector search applications and in automatic content tagging systems.