一通预测目标与立场，模型更小性能略降但部署更易。
From Two Passes to One: Compact and Efficient Target-Stance Extraction
- 用单次前向传播同时预测目标和立场，取代传统两步流程。
- 参数量减半，性能仅下降4-7个F1点，目标预测准确率提升至少6点。
- 保留真实文本中的目标词能显著提升效果，适合实际应用落地。
目标立场抽取（TSE）旨在预测作者言论的目标（或主题）及其立场。现有方法采用两阶段神经模型流水线：先识别目标，再判断立场。本文提出一种单次前向传播的联合架构，可在一次计算中完成两项任务，将可训练参数减少近50%，性能仅损失4-7个F1点。我们进一步证明，标准的目标清洗做法会人为压制目标预测准确率。在真实场景中保留显式目标提及，能在目标分类与目标生成两种设置下，使F1值提升至少6点。这一改进使得TSE更容易集成到如舆情追踪等下游应用中。
原文摘要 · Abstract (English)
Target-Stance Extraction (TSE) is the task of predicting both the target (or topic) of an author's writing and the author's stance toward it. Existing approaches to TSE use a sequential pipeline of two separate neural models: one to identify the target and another to determine the stance. We present a one-pass, joint architecture that predicts both in a single forward pass, reducing trainable parameters by nearly 50% with only a 4-7 F1 point tradeoff in performance. We further demonstrate that standard target-scrubbing practices artificially suppress target prediction accuracy. Retaining explicit target mentions, as in real-world deployments, improves F1 by at least 6 points across both target classification and target generation settings. These improvements allow for significantly easier integration of TSE in downstream applications such as public opinion tracking.
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。