2026

LongCat-Flash-Prover: Advancing Native Formal Reasoning via Agentic Tool-Integrated Reinforcement Learning

MLC Team

arXiv preprint 2026

LongCat-Flash-Prover advances native formal reasoning capabilities through agentic tool-integrated reinforcement learning.

LongCat-Flash-Prover: Advancing Native Formal Reasoning via Agentic Tool-Integrated Reinforcement Learning

MLC Team

arXiv preprint 2026

LongCat-Flash-Prover advances native formal reasoning capabilities through agentic tool-integrated reinforcement learning.

Longcat-Flash-Thinking-2601 Technical Report

MLC Team

arXiv preprint 2026

Technical report of Longcat-Flash-Thinking-2601, an advanced reasoning model by the MLC Team.

Longcat-Flash-Thinking-2601 Technical Report

MLC Team

arXiv preprint 2026

Technical report of Longcat-Flash-Thinking-2601, an advanced reasoning model by the MLC Team.

2025

Introducing LongCat-Flash-Thinking: A Technical Report

MLC Team

arXiv preprint 2025

Technical report introducing LongCat-Flash-Thinking, extending LongCat-Flash with enhanced reasoning capabilities.

Introducing LongCat-Flash-Thinking: A Technical Report

MLC Team

arXiv preprint 2025

Technical report introducing LongCat-Flash-Thinking, extending LongCat-Flash with enhanced reasoning capabilities.

LongCat-Flash Technical Report

MLC Team

arXiv preprint 2025

Technical report of LongCat-Flash, a large language model developed by the MLC Team.

LongCat-Flash Technical Report

MLC Team

arXiv preprint 2025

Technical report of LongCat-Flash, a large language model developed by the MLC Team.

Harnessing the Power of Reinforcement Learning for Language-Model-Based Information Retriever via Query-Document Co-Augmentation

Jingming Liu, Yumeng Li, Wei Shi, Yao-Xiang Ding, Hui Su, Kun Zhou

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD) 2026

We propose a reinforcement learning approach for language-model-based information retrieval through query-document co-augmentation.

Harnessing the Power of Reinforcement Learning for Language-Model-Based Information Retriever via Query-Document Co-Augmentation

Jingming Liu, Yumeng Li, Wei Shi, Yao-Xiang Ding, Hui Su, Kun Zhou

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD) 2026

We propose a reinforcement learning approach for language-model-based information retrieval through query-document co-augmentation.

2021

Entity Enhancement for Implicit Discourse Relation Classification in the Biomedical Domain

Wei Shi, Vera Demberg

Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP) 2021

We propose an entity-enhanced model for implicit discourse relation classification in the biomedical domain, leveraging entity information to improve relation detection.

Entity Enhancement for Implicit Discourse Relation Classification in the Biomedical Domain

Wei Shi, Vera Demberg

Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP) 2021

We propose an entity-enhanced model for implicit discourse relation classification in the biomedical domain, leveraging entity information to improve relation detection.

2020

Addressing the Data Bottleneck in Implicit Discourse Relation Classification

Wei Shi

PhD Thesis, Saarland University 2020

This dissertation addresses the data bottleneck problem in implicit discourse relation classification through cross-lingual and transfer learning approaches.

Addressing the Data Bottleneck in Implicit Discourse Relation Classification

Wei Shi

PhD Thesis, Saarland University 2020

This dissertation addresses the data bottleneck problem in implicit discourse relation classification through cross-lingual and transfer learning approaches.

2019

Next Sentence Prediction helps Implicit Discourse Relation Classification within and across Domains

Wei Shi, Vera Demberg

Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) 2019

We show that next sentence prediction, as used in BERT pre-training, can effectively help implicit discourse relation classification both within and across domains.

Next Sentence Prediction helps Implicit Discourse Relation Classification within and across Domains

Wei Shi, Vera Demberg

Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) 2019

We show that next sentence prediction, as used in BERT pre-training, can effectively help implicit discourse relation classification both within and across domains.

A Hybrid Model for Globally Coherent Story Generation

Fangzhou Zhai, Vera Demberg, Pavel Shkadzko, Wei Shi, Asad Sayeed

Workshop on Narrative Understanding (StoryNLP@ACL) 2019

We propose a hybrid model that combines a language model with a discourse-level planning component to generate globally coherent stories.

A Hybrid Model for Globally Coherent Story Generation

Fangzhou Zhai, Vera Demberg, Pavel Shkadzko, Wei Shi, Asad Sayeed

Workshop on Narrative Understanding (StoryNLP@ACL) 2019

We propose a hybrid model that combines a language model with a discourse-level planning component to generate globally coherent stories.

Acquiring Annotated Data with Cross-lingual Explicitation for Implicit Discourse Relation Classification

Wei Shi, Frances Yung, Vera Demberg

Workshop on Discourse Relation Parsing and Treebanking (DISRPT@NAACL) 2019

We explore cross-lingual explicitation as a method for acquiring annotated data for implicit discourse relation classification.

Acquiring Annotated Data with Cross-lingual Explicitation for Implicit Discourse Relation Classification

Wei Shi, Frances Yung, Vera Demberg

Workshop on Discourse Relation Parsing and Treebanking (DISRPT@NAACL) 2019

We explore cross-lingual explicitation as a method for acquiring annotated data for implicit discourse relation classification.

Learning to Explicitate Connectives with Seq2Seq Network for Implicit Discourse Relation Classification

Wei Shi, Vera Demberg

13th International Conference on Computational Semantics (IWCS) 2019

We propose a Seq2Seq approach to explicitate implicit discourse connectives, and use the generated connectives to improve implicit discourse relation classification.

Learning to Explicitate Connectives with Seq2Seq Network for Implicit Discourse Relation Classification

Wei Shi, Vera Demberg

13th International Conference on Computational Semantics (IWCS) 2019

We propose a Seq2Seq approach to explicitate implicit discourse connectives, and use the generated connectives to improve implicit discourse relation classification.

2018

'ONLY' increases Expectations for Causal Coherence Relations

Katharina Spalek, Birte Bergmann, Wei Shi, Vera Demberg

Discourse Expectations: Theoretical, Experimental, and Computational Perspectives 2019

We investigate how the focus particle 'only' influences expectations for causal coherence relations in discourse processing.

'ONLY' increases Expectations for Causal Coherence Relations

Katharina Spalek, Birte Bergmann, Wei Shi, Vera Demberg

Discourse Expectations: Theoretical, Experimental, and Computational Perspectives 2019

We investigate how the focus particle 'only' influences expectations for causal coherence relations in discourse processing.

2017

Using Explicit Discourse Relation Connectives in Translation for Implicit Discourse Relation Classification

Wei Shi, Frances Yung, Raphael Rubino, Vera Demberg

8th International Joint Conference on Natural Language Processing (IJCNLP) 2017

We leverage explicit discourse connectives that appear in translations to obtain additional training signal for implicit discourse relation classification.

Using Explicit Discourse Relation Connectives in Translation for Implicit Discourse Relation Classification

Wei Shi, Frances Yung, Raphael Rubino, Vera Demberg

8th International Joint Conference on Natural Language Processing (IJCNLP) 2017

We leverage explicit discourse connectives that appear in translations to obtain additional training signal for implicit discourse relation classification.

On the Need of Cross Validation for Discourse Relation Classification

Wei Shi, Vera Demberg

15th Conference of the European Chapter of the Association for Computational Linguistics (EACL) 2017

We demonstrate the need for cross-validation in discourse relation classification, showing that standard train/test splits can lead to unreliable evaluation.

On the Need of Cross Validation for Discourse Relation Classification

Wei Shi, Vera Demberg

15th Conference of the European Chapter of the Association for Computational Linguistics (EACL) 2017

We demonstrate the need for cross-validation in discourse relation classification, showing that standard train/test splits can lead to unreliable evaluation.

2016

Attention-based Bidirectional Long Short-term Memory Networks for Relation Classification

Peng Zhou, Wei Shi, Jun Tian, Zhenyu Qi, Bingchen Li, Hongwei Hao, Bo Xu

54th Annual Meeting of the Association for Computational Linguistics (ACL) 2016

We propose an attention-based bidirectional LSTM model for relation classification, which can capture the most important semantic information in a sentence.

Attention-based Bidirectional Long Short-term Memory Networks for Relation Classification

Peng Zhou, Wei Shi, Jun Tian, Zhenyu Qi, Bingchen Li, Hongwei Hao, Bo Xu

54th Annual Meeting of the Association for Computational Linguistics (ACL) 2016

We propose an attention-based bidirectional LSTM model for relation classification, which can capture the most important semantic information in a sentence.