IJCAI

Self-Classification Enhancement and Correction for Weakly Supervised Object Detection

自分类增强与修正用于弱监督目标检测

Yufei Yin, Lechao Cheng, Wengang Zhou, Jiajun Deng, Zhou Yu, Houqiang Li

AI总结 本文提出一种新的弱监督目标检测框架，通过自分类增强模块和修正算法解决分类模糊性和误分类问题，实验表明其在VOC数据集上表现优异。

Comments Accepted by IJCAI 2025

Formal Abductive Explanations for Navigating Mental Health Help-Seeking and Diversity in Tech Workplaces

形式归纳解释用于导航心理健康求助与科技职场多样性

Belona Sonna, Alain Momo, Alban Grastien

AI总结 本文提出形式归纳解释框架，用于系统揭示AI对科技职场心理健康求助预测的合理性。通过计算模型输出的严谨解释，支持选择适合不同心理科目的模型，并确保伦理稳健的救济计划。同时，明确考察性别等敏感属性对模型决策的影响，以支持公平性评估。

Comments Appeared in the Proceedings of the Empowering Women of Colour in AI-Driven Mental Health Research at IJCAI 2025

Localizing Before Answering: A Hallucination Evaluation Benchmark for Grounded Medical Multimodal LLMs

先定位后回答：一种用于基于地面的医学多模态大语言模型的幻觉评估基准

Dung Nguyen, Minh Khoi Ho, Huy Ta, Thanh Tam Nguyen, Qi Chen, Kumar Rav, Quy Duong Dang, Satwik Ramchandre, Son Lam Phung, Zhibin Liao, Minh-Son To, Johan Verjans, Phi Le Nguyen, Vu Minh Hieu Phan

AI总结 本文提出HEAL-MedVQA基准，通过创新评估协议和67K VQA数据集，评估医学多模态模型的定位能力与幻觉鲁棒性，提出LobA框架提升视觉推理能力，实验结果表明其在挑战性基准中优于现有生物医学LMMs。

Comments Accepted at Joint Conference on Artificial Intelligence (IJCAI) 2025