Home
From Oct. 5, 2020, I joined NVIDIA Japan as a Senior Solution Architect - Deep Learning.
My new working email box is: xianchaow at nvidia.com, or you can find me from wuxianchao at gmail.com. Currently, I am focusing on the following topics:
1. Finance + NVIDIA AI + GPU, such as NLP + Finance (research report generation, summarization, stock market understanding, and so on);
2. General NLP, such as Megatron BERT/GPT and TensorRT for model parallel 512+ GPU training/inferencing of extremely large multi-lingual language models;
3. Jarvis [Riva]- Conversational AI with NeMo, Triton and so on for multi-modal voice-to-voice and video-based human-chatbot communication.
Since I support generally in Japanese market, if you face any issues of using GPU or you are interested in our open-source solutions, do share with us!
---------I am the separator ----------
From early Aug. 2015 to Oct. 4, 2020, I joined Microsoft Japan, Bing Team as a Software Development Engineer
(Senior Software Engineer and A.I. Scientist; from 2017.03).
My old working email box was: xiancwu at microsoft.com [useless NOW]
From Nov. 2018, I began to focus on AI creations, in fields include NLP, image, music, video and so on.
The central target is to drive AI creation oriented applications and user-oriented products.
In addition, I am doing research on machine reading comprehension.
Well, I just want to forget everything related to my working experience during that several years...
----separator----
I was a senior R&D in Baidu Inc. working in Tokyo, from April 2012 to July 2015 (3 years and 4 months). The memory of working at Baidu was nice.
I obtained my Ph.D. degree (Computer Science) from the Tsujii Lab. in The University of Tokyo in Mar. 2010.
I obtained my MBA degree (finance strategy) from the Honda Lab. in Hitotsubashi University in Mar. 2020.
Before joining Baidu, I worked as a Research Associate in NTT Communication Science Laboratories for one year,
and as a Post-doc researcher in The University of Tokyo for one year.
My main research topic was Statistical Machine Translation.
I have finished my work on Baidu Japanese IME for Windows and Android (2012.04 to 2013.05) ,
Chinese syntactic/semantic parsing, especially shift-reduce dependency parsing (2013.06 to 2014.07)
and semantic role labeling (SRL, 2013.11 to 2014.07).
From 2014.07 to 2015.07, I have focused on improving the relevance of query/pages in information retrieval,
using large-scale click-through data, deep learning techniques, neural statistical machine translation models,
as well as syntactic/semantic structures.
My personal email address = (X=wu and Y=xianchao) for XY AT gMAILDOTcom
Check my open source Akamon System for "HPSG forest to string Statistical Machine Translation"!.
My name in Chinese characters: 吴(Wu) 先超(Xianchao)
My name in Japanese Kanji: 呉 先超(ご せんちょう -
alike Gao-Shanqiao/高山桥 in Chinese pronunciation)
Find my CV here. (last updated 2020.06.08)
I have published more than 40 papers/articles, and applied more than 90 (written in Japanese/Chinese/English) patents.
A relatively full list of my patents can be found HERE (for 46 patents in English/Chinese, mainly when I worked in Baidu)
and HERE (for 6 patents in Japanese, mainly when I worked in NTT).