2024

• ChatQA 2: Bridging the Gap to Proprietary LLMs in Long Context and RAG Capabilities [arxiv]

  • Peng Xu, Wei Ping, Xianchao Wu, Zihan Liu, Mohammad Shoeybi, and Bryan Catanzaro.

2023

• SteerLM: Attribute Conditioned SFTas an (User-Steerable) Alternative to RLHF [arxiv]

  • Yi Dong, Zhilin Wang,Makesh Narsimhan Sreedhar, Xianchao Wu and Oleksii Kuchaiev .

  • In EMNLP 2023 Findings.

• Retrieval meets Long Context Large Language Models [arxiv]

  • Peng Xu, Wei Ping, Xianchao Wu, Lawrence McAfee, Chen Zhu, Zihan Liu, Sandeep Subramanian, Evelina Bakhturina, Mohammad Shoeybi and Bryan Catanzaro.

  • To Appear in ICLR 2024.

• Duplex Diffusion Models Improve Speech-to-Speech Translation [arxiv]

  • Xianchao Wu.

  • In ACL 2023 Findings.

• Enhancing Unsupervised Speech Recognition with Diffusion GANs 

  • Xianchao Wu.

  • In ICASSP 2023.

2022

• Creative Painting with Latent Diffusion Models [arxiv version] [coling 2022 workshop version] [bib]

  • Xianchao Wu.

  • In COLING 2022, Workshop on When Creative AI Meets Conversational AI.

• Attention Enhanced Citrinet for Speech Recognition

  • Xianchao Wu.

  • In  Interspeech 2022. Tue-P-VR-4-4 Tuesday, September 20, 13:30-15:30(KST), Virtual Poster: Neural Transducers, Streaming ASR and Novel ASR Models

  • Code: https://github.com/Xianchao-Wu/nemo_bidecoder 

• Deep Sparse Conformer for Speech Recognition

  • Xianchao Wu.

  • In Interspeech 2022. Tue-P-VR-4-4 Tuesday, September 20, 13:30-15:30(KST), Virtual Poster: Neural Transducers, Streaming ASR and Novel ASR Models

  • Code: https://github.com/Xianchao-Wu/wenet-deep-sparse-conformer 

2021

• NVJPFSI at FinCausal 2021 Span-based Causality Extraction Task

  • Xianchao Wu.

  • In Proceedings of the 3rd Financial Narrative Processing Workshop .

• When Creative AI Meets Conversational AI.

  • A Video Talk to NVIDIA GTC 2021 (April).

  • PPT/PDF is HERE.

• Multi-modal based conversational AI empowered by GPU.

  • Invited talk to IoT-SNAP-2021.

  • The PDF can also be find HERE.

• When Creative AI Meets Conversational AI. 

  • Workshop organizer. Joint held with Japan NLP 2021. 

• FinMegatron: Large Financial Domain Language Models 

  • Xianchao Wu.

  • In SIG-FIN-026.

2020

• Event-Driven Learning of Systematic Behaviours in Stock Markets

  • Xianchao Wu.

  • In EMNLP 2020 Findings.

  • some old code/test sets can be find in: URL: https://pan.baidu.com/s/1EqCu49f1LgjLmYF7EecYjQ  PIN code: esf8  

• Transformer-XL Based Music Generation with Multiple Sequences of Time-valued Notes

  • Xianchao Wu, Chengyuan Wang, Qinying Lei.

  • Arxiv, 2020.

  • Demo music generated: https://www.youtube.com/watch?v=yNFMxlhgAmo

2019

• Learning-to-Explain: Recommendation Reason Determination Through Q20 Gaming

  • Xianchao Wu.

  • SIGIR 2019's EARS 2019. July, Paris, France, 2019.

• AI 歌手りんな：ユーザ歌唱や楽譜を入力とする歌声合成システム 

  • 沢田 慶，坪井 一菜，Wu Xianchao，Chen Zhan(MSD)，法野 行哉，橋本 佳，大浦 圭一郎，南角 吉彦，徳田 恵一(NITech)

  • In 日本音響学会春季研究発表会 March, 2019

• Learning-to-Suggest: Product Recommendation via Several Questions

  • Xianchao Wu.

  • In Proceedings of Japan NLP 2019. Nagoya, Japan. March, 2018.

2018

• Wrote 15+ patents.

• Playing 20 Question Game with Policy-Based Reinforcement Learning

  • Huang Hu, Xianchao Wu, Bingfeng Luo, Chongyang Tao, Can Xu, Wei Wu, and Zhan Chen

  • In Proceedings of EMNLP 2018. Oct-Nov, 2018.

• Dialog Generation Using Multi-turn Reasoning NeuralNetworks

  • Xianchao Wu, Ander Martinez and Momo Klyen.

  • In Proceedings of NAACL-HLT 2018. pages 2049-2059. USA. June, 2018.

• Q20: Rinna Riddles Your Mind by Asking 20 Questions

  • Xianchao Wu, Huang Hu, Momo Klyen, Kyohei Tomita, Zhan Chen.

  • In Proceedings of Japan NLP 2018. Okayama, Japan. March, 2018

• Evaluating Rinna's Mind-reading Feature by Self-playing

  • Xianchao Wu, Huang Hu.

  • In Proceedings of Japan NLP 2018. Okayama, Japan. March, 2018.

2017

• Wrote 15+ patents.

• Fine-Grained Sentiment Analysis with 32 Dimensions

  • Xianchao Wu, Hang Tong and Momo Klyen.

  • In Proceedings of 21th International Conference on Asian Language Processing (IALP). Dec, 2017, Singapore.

• Rinna’s CharBox: from Pure Chat to Product Recommendation

  • Xianchao Wu, Keizo Fujiwara, Katsuya Iida, Kyohei Tomita, Rica Nakajima.

  • In Proceedings of 81th Language/Voice Understanding and Dialog Processing Research Symposium (第81回言語・音声理解と対話処理研究会). Waseda, Japan.

• Sentiment Analysis with Eight Dimensions for Emotional Chatbots

  • Xianchao Wu, Yuichiro Kikura, Momo Klyen, Zhan Chen.

  • In Proceedings of Japan NLP 2017. Tsukuba, Japan.

• Haiku Generation Using Deep Neural Networks

  • Xianchao Wu, Momo Klyen, Kazushige Ito, Zhan Chen.

  • In Proceedings of Japan NLP 2017. Tsukuba, Japan.

2016

• りんな：女子高生人工知能

  • Xianchao Wu, Kazushige Ito, Katsuya Iida, Kazuna Tsuboi, Momo Klyen.

  • In Proceedings of Japan NLP 2016. Sendai, Japan.

2015

      Wrote 10+ patents

2014

      Wrote 20+ patents

2013

• Generalization of Words for Chinese Dependency Parsing

  • Xianchao Wu, Jie Zhou, Yu Sun, Zhanyi Liu, Dianhai Yu, Hua Wu, and Haifeng Wang.

  • In Proceedings of IWPT 2013. Nara, Japan.

  • Using the Web to Train a Mobile Device Oriented Japanese Input Method Editor <bib>

  • Xianchao Wu, Rixin Xiao, and Xiaoxin Chen

  • In Proceedings of IJCNLP 2013. Nagoya, Japan.

• Mining Japanese Compound Words and Their Pronunciations from Web Pages and Tweets <bib>

  • Xianchao Wu.

  • In Proceedings of IJCNLP 2013. Nagoya, Japan.

• Using a Chunk-based Dependency  Parser to Mine Compound Words from Tweets

  • Xianchao Wu.

    • In Proceedings of Japan Natural Language Processing 2013.

• Syntax-based Post-ordering for Efficient Japanese-to-English Translation

  • • Katsuhito Sudoh, Xianchao Wu, Kevin Duh, Hajime Tsukada, and Masaaki Nagata

    • ACM Trans. on Asian Language Information Processing (TALIP). 

    • This journal paper is partially based on Sudoh et al.'s MT summit 2011 paper.

2012

• Using Collocations and K-means Clustering to Improve the N-pos Model for Japanese IME  <bib>

  • • Long Chen, Xianchao Wu, and Jingzhou He

    • In Proceedings of Second Workshop on Advances in Text Input Methods (WTIM 2), collocated with COLING 2012.

    • The Baidu Japanese IME system can be FREELY downloaded here: http://ime.baidu.jp/type/?source=pstop

• Akamon: An Open Source Toolkit for Tree/Forest-Based Statistical Machine Translation. <bib>

  • • Xianchao Wu, Takuya Matsuzaki, and Jun'ichi Tsujii.  In proceedings of ACL 2012, demo session. 

    • Code is here:  http://code.google.com/p/akamon-forest-to-string-translation/  

• A Comparative Study of Target Dependency Structures for Statistical Machine Translation. <bib>

  • • Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada, and Masaaki Nagata. In proceedings of ACL 2012, short paper

• Learning to Translate with Multiple Objectives <bib>

  • • Kevin Duh, Katsuhito Sudoh, Xianchao Wu, Hajime Tsukada, and Masaaki Nagata. In proceedings of ACL 2012, long paper.

• Head Finalization Reordering for Chinese-to-Japanese Machine Translation <bib>

• • Dan Han, Katsuhito Sudoh, Xianchao Wu, Kevin Duh, Hajime Tsukada, and Masaaki Nagata. In proceedings of SSST-2012.

• An Improvement to the Predicate-Argument Structure Based Pre-ordering Approach for Statistical Machine Translation

  • • Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada, and Masaaki Nagata. In Proceedings of Japan NLP 2012. March, 2012. Hiroshima, Japan.

• Syntactic Based Reordering Rules for Chinese-to-Japanese Machine Translation

  • • Dan Han, Katsuhito Sudoh, Xianchao Wu, Kevin Duh, Hajime Tsukada, and Masaaki Nagata. In Proceedings of Japan NLP 2012. March, 2012. Hiroshima, Japan.

2011

• SMT Systems in the University of Tokyo for NTCIR-9 PatentMT

  • • Xianchao Wu, Takuya Matsuzaki and Jun'ichi Tsujii. In Proceedings of NTCIR-9. December, 2011. Tokyo, Japan. <PDF>

• NTT-UT Statistical Machine Translation in NTCIR-9 PatentMT

  • • Katsuhito Sudoh, Kevin Duh, Hajime Tsukada, Masaaki Nagata, Xianchao Wu, Takuya Matsuzaki and Jun'ichi Tsujii. In Proceedings of NTCIR-9. December, 2011. Tokyo, Japan.1st-rank in E2J (first time better than rule-based translation systems) and 1st-rank in J2E SMT systems<PDF>

• Statistical Machine Translation Using Large-Scale Lexicon and Deep Syntactic Structures

  • • Xianchao Wu. In Quick Report on Doctoral Theses Recommended by IPSG SIGs. Journal of Information Processing. 52(10).

• Technical Report of NTT for the 7th China Workshop on Machine Translation

  • • Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada and Masaaki Nagata. 2011. In Proceedings of 7th China Workshop on Machine Translation. <PDF>

• Extracting Pre-ordering Rules from Chunk-based Dependency Trees for Japanese-to-English Translation

  • • Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada and Masaaki Nagata. 2011. In Proceedings of MT summit 2011< PDF >

• Post-ordering in Statistical Machine Translation

  • • Katsuhito Sudoh, Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada and Masaaki Nagata. 2011. In Proceedings of MT summit 2011<PDF>

• Extracting Pre-ordering Rules from Predicate-Argument Structures

  • • Xianchao Wu, Katsuhito Sudoh, Kevin Duh, Hajime Tsukada and Masaaki Nagata. 2011. In Proceedings of IJCNLP 2011 <PDF> Finalist of Best Paper Award

• Generalized Minimum Bayes Risk System Combination

  • • Kevin Duh, Katsuhito Sudoh, Xianchao Wu, Hajime Tsukada and Masaaki Nagata. 2011. In Proceedings of IJCNLP 2011 <PDF>

• A Multi-Objective Approach Based on Genetic Algorithm for Multi-Model Line Operation Planning Considering Difference in Worker Ability

  • • Jiahua Weng, Xianchao Wu, Hisashi Onari. 2011. In Proceedings of 21st International Conference on Production Research. Stuttgart, Germany

• Effective Use of Function Words for Rule Generalization in Forest-Based Translation

  • • Xianchao Wu, Takuya Matsuzaki, and Jun'ichi Tsujii. 2011. In Proceedings of ACL-HLT 2011 <PDF>

• A Term Translation System Using Hierarchical Phrases and Morphemes

  • • Xianchao Wu and Jun'ichi Tsujii. 2011. In Proceedings of Japan NLP 2011. <PDF>

2010

• Fine-grained Tree-to-String Translation Rule Extraction

  • • Xianchao Wu, Takuya Matsuzaki, and Jun'ichi Tsujii. 2010. In Proceedings of ACL 2010, 10 pages. <PDF> <bib> <slides>

• Improve Syntax-based Translation with Deep Syntactic Structures

  • • Xianchao Wu, Takuya Matsuzaki, and Jun'ichi Tsujii. 2010. Machine Translation, Volume 24, Number 2, 141-157. DOI: 10.1007/s10590-010-9081-6 <Draft>

• • Statistical Machine Translation Using Large-Scale Lexicon and Deep Syntactic Structures <PDF> <bib>

    • • Xianchao Wu (PhD thesis draft, all rights reserved)

  • Deep Syntactic Structures for String-to-Tree Translation (Revised) <PDF> <bib>

    • • Xianchao Wu and Jun'ichi Tsujii. 2010. In Proc. of Japan NLP.

2009

• • The UOT System: Improve String-to-Tree Translation Using Head-Driven Phrasal Structure Grammar and Predicate-Argument Structures <PDF> <bib>

    • • Xianchao Wu, Takuya Matsuzaki, Naoaki Okazaki, Yusuke Miyao, and Jun'ichi Tsujii. 2009. In Proceedings of IWSLT 2009, pages 99-106, Tokyo, Japan.

  • Semi-Supervised Lexicon Mining from Parenthetical Expressions in Monolingual Web Pages <PDF> <bib>

    • • Xianchao Wu, Naoaki Okazaki, and Jun'ichi Tsujii. In Processings of NAACL-HLT 2009, 9 pages.

  • Self-training for mining parenthetical translations in monolingual web pages <bib>

    • • Xianchao Wu, Naoaki Okazaki, and Jun'ichi Tsujii. In Proceedings of Japan NLP 2009, 4 pages.

2008

• • Improving English-to-Chinese Translation for Technical Terms Using Morphological Information <PDF> <bib>

    • • Xianchao Wu, Naoaki Okazaki, Takashi Tsunakawa and Jun'ichi Tsujii. In Proceedings of AMTA 2008, 10 pages.

2006

• • Research on configuration & integration based on generic product structure (in Chinese) <PDF> <bib>

    • Xianchao Wu, Xiaofeng Lv, and Jigui Sun. Computer Integrated Manufacturing Systems (2006). 12(6):868-875.

---

Patents (partial, check my resume for a full list)

English and Chinese

TitleCtrPubDateInt.ClassAppl.NoApplicantInventor

1.20160370959 METHOD AND DEVICE FOR UPDATING INPUT METHOD SYSTEM, COMPUTER STORAGE MEDIUM, AND DEVICEUS22.12.2016