Multi-turn Retrieval-Augmented Generation faces structural challenges that go beyond single-turn retrieval and fusion. Context-dependent queries, cross-turn evidence accumulation, and uncertain answerability jointly affect retrieval quality and generation reliability. We propose a structured control framework that formulates multi-turn RAG as a regulated reasoning process rather than a loosely coupled pipeline. The system first performs evidence and context structuring, extracting atomic facts strictly grounded in reference passages while reconstructing a self-contained query from dialogue history. It then conducts decision-conditioned generation, where explicit control signals regarding question intent, dialogue dependency, and answerability govern response feasibility, scope, and organization. By separating structural decision making from surface realization, the framework enforces consistent information flow across stages and reduces hallucination.Experiments on SemEval-2026 Task 8 show that our approach achieves strong faithfulness and stable overall performance, ranking 17/26 on Task B (generation, H=0.6333).

This paper describes TenTrans large-scale multilingual machine translation system for WMT 2021. We participate in the Small Track 2 in five South East Asian languages, thirty directions: Javanese, Indonesian, Malay, Tagalog, Tamil, English. We mainly utilized forward/back-translation, in-domain data selection, knowledge distillation, and gradual fine-tuning from the pre-trained model FLORES-101. We find that forward/back-translation significantly improves the translation results, data selection and gradual fine-tuning are particularly effective during adapting domain, while knowledge distillation brings slight performance improvement. Also, model averaging is used to further improve the translation performance based on these systems. Our final system achieves an average BLEU score of 28.89 across thirty directions on the test set.

This paper describes TenTrans’ submission to WMT21 Multilingual Low-Resource Translation shared task for the Romance language pairs. This task focuses on improving translation quality from Catalan to Occitan, Romanian and Italian, with the assistance of related high-resource languages. We mainly utilize back-translation, pivot-based methods, multilingual models, pre-trained model fine-tuning, and in-domain knowledge transfer to improve the translation quality. On the test set, our best-submitted system achieves an average of 43.45 case-sensitive BLEU scores across all low-resource pairs. Our data, code, and pre-trained models used in this work are available in TenTrans evaluation examples.

Natural language inference (NLI) is a central problem in language understanding. End-to-end artificial neural networks have reached state-of-the-art performance in NLI field recently. In this paper, we propose Character-level Intra Attention Network (CIAN) for the NLI task. In our model, we use the character-level convolutional network to replace the standard word embedding layer, and we use the intra attention to capture the intra-sentence semantics. The proposed CIAN model provides improved results based on a newly published MNLI corpus.