Zhentao Xia

Yongqi Fan, Yuxiang Chu, Zhentao Xia et al.

In the RAG paradigm, document ranking determines the evidence available to downstream generators. Through controlled analysis, we identify two phenomena underexplored by existing rankers: (i) downstream response quality depends not only on relevance but also on the composition and ordering of selected documents, and (ii) such preferences differ systematically across generators. However, existing rankers are trained purely on query--document relevance, leaving both phenomena unmodeled. To close this gap, we construct \textbf{PRISM}, a bilingual preference-aligned dataset built through a four-stage pipeline that compresses the combinatorial subset-and-ordering space by roughly four orders of magnitude and produces response-quality preference supervision conditioned on seven downstream generators. On a 13k-query subset of PRISM, we train \textbf{Rank4Gen}, a generator-aware ranker that performs joint document set selection and ordering. Experiments on five challenging RAG benchmarks show that Rank4Gen improves downstream QA quality on most evaluated generators, with per-generator F1 gains of up to $+2.08$ over the strongest set-selection baseline. Code is available at https://github.com/JOHNNY-fans/Rank4Gen.

Zhentao Xia, Likai Wang, Weiguang Qu et al.

In this paper, we describe the details of the neural dependency parser sub-mitted by our team to the NLPCC 2019 Shared Task of Semi-supervised do-main adaptation subtask on Cross-domain Dependency Parsing. Our system is based on the stack-pointer networks(STACKPTR). Considering the im-portance of context, we utilize self-attention mechanism for the representa-tion vectors to capture the meaning of words. In addition, to adapt three dif-ferent domains, we utilize neural network based deep transfer learning which transfers the pre-trained partial network in the source domain to be a part of deep neural network in the three target domains (product comments, product blogs and web fiction) respectively. Results on the three target domains demonstrate that our model performs competitively.