The PerAnsSumm Shared Task at CL4Health@NAACL 2025 focused on Perspective-Aware Summarization of Healthcare Q/A forums, requiring participants to extract and summarize spans based on predefined perspective categories. Our approach leveraged LLM-based zero-shot prompting enhanced by semantically-similar In-Context Learning (ICL) examples. Using Qwen-Turbo with 20 exemplar samples retrieved through NV-Embed-v2 embeddings, we achieved a mean score of 0.58 on Task A (span identification) and Task B (summarization) mean scores of 0.36 in Relevance and 0.28 in Factuality, finishing 12th on the final leaderboard. Notably, our system achieved higher precision in strict matching (0.20) than the top-performing system, demonstrating the effectiveness of our post-processing techniques. In this paper, we detail our ICL approach for adapting Large Language Models to Perspective-Aware Medical Summarization, analyze the improvements across development iterations, and finally discuss both the limitations of the current evaluation framework and future challenges in modeling this task. We release our code for reproducibility.

Citation Prediction, estimating whether paper a cites paper b, is particularly interesting in a forecasting setting where the model is trained on papers published before time t, and evaluated on papers published after h, where h is the forecast horizon. Performance improves with t (larger training sets) and degrades with h (longer forecast horizons). The trade-off between edge-based methods and node-based methods depends on t. Because edges grow faster than nodes, larger training sets favor edge-based methods.We introduce a new forecast-based Citation Prediction benchmark of 3 million papers to quantify these trends.Our benchmark shows that desirable policies for combining edge- and node-based methods depend on h and t.We release our benchmark, evaluation scripts, and embeddings.

Visual Question Answering (VQA) methods aim at leveraging visual input to answer questions that may require complex reasoning over entities. Current models are trained on labelled data that may be insufficient to learn complex knowledge representations. In this paper, we propose a new method to enhance the reasoning capabilities of a multi-modal pretrained model (Vision+Language BERT) by integrating facts extracted from an external knowledge base. Evaluation on the KVQA dataset benchmark demonstrates that our method outperforms competitive baselines by 19%, achieving new state-of-the-art results. We also perform an extensive analysis highlighting the limitations of our best performing model through an ablation study.

The CogNLP-Sheffield submissions to the CMCL 2021 Shared Task examine the value of a variety of cognitively and linguistically inspired features for predicting eye tracking patterns, as both standalone model inputs and as supplements to contextual word embeddings (XLNet). Surprisingly, the smaller pre-trained model (XLNet-base) outperforms the larger (XLNet-large), and despite evidence that multi-word expressions (MWEs) provide cognitive processing advantages, MWE features provide little benefit to either model.