The HLTCOE Evaluation team participated in TREC VQA's Answer Generation (AG) task, for which we developed a listwise learning framework that aims to improve semantic precision and ranking consistency in answer generation. Given a video-question pair, a base multimodal model first generates multiple candidate answers, which are then reranked using a model trained with a novel Masked Pointer Cross-Entropy Loss with Rank Weights. This objective integrates pointer-based candidate selection, rank-dependent weighting, and masked cross-entropy under vocabulary restriction, enabling stable and interpretable listwise optimization. By bridging generative modeling with discriminative ranking, our method produces coherent, fine-grained answer lists. Experiments reveal consistent gains in accuracy and ranking stability, especially for questions requiring temporal reasoning and semantic disambiguation.
@article{arxiv.2512.07738,
title = {HLTCOE Evaluation Team at TREC 2025: VQA Track},
author = {Dengjia Zhang and Charles Weng and Katherine Guerrerio and Yi Lu and Kenton Murray and Alexander Martin and Reno Kriz and Benjamin Van Durme},
journal= {arXiv preprint arXiv:2512.07738},
year = {2025}
}