通过置信度阈值控制,让视频问答模型在不确定时主动放弃预测,提升可靠性。
Explicit Abstention Knobs for Predictable Reliability in Video Question Answering
- 用置信度阈值实现模型在不确定时主动放弃回答
- 在原分布下可平滑调节错误率与覆盖范围的权衡
- 在分布外数据上仍保持可靠控制,适合高风险场景
高风险场景中部署视觉语言模型(VLMs)需要选择性预测,即在不确定时放弃输出以避免严重错误。我们研究了基于置信度的放弃机制是否能在视频问答任务中提供可靠的错误率控制,并且该控制在分布偏移下是否依然稳健。基于NExT-QA数据集和Gemini 2.0 Flash模型,我们得出两个结论:第一,在分布内,置信度阈值能实现机制上的可控性,调整阈值ε可产生平滑的风险-覆盖率权衡,将错误率从35.6%降至10.8%;第二,该控制在分布外仍具鲁棒性,说明其具备实际部署潜力。
原文摘要 · Abstract (English)
High-stakes deployment of vision-language models (VLMs) requires selective prediction, where systems abstain when uncertain rather than risk costly errors. We investigate whether confidence-based abstention provides reliable control over error rates in video question answering, and whether that control remains robust under distribution shift. Using NExT-QA and Gemini 2.0 Flash, we establish two findings. First, confidence thresholding provides mechanistic control in-distribution. Sweeping threshold epsilon produces smooth risk-coverage tradeoffs, reducing error rates f
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。