wonderseen / commonsense-modeling Goto Github PK

I am an NLP algorithm engineer graduated from Xiamen University with bachelor degree (2015-2019) and Tianjin University with master degree (2019-2022).

My research interests include:

• 🔭 clustering analysis (fuzzy clustering theory and linguistic clustering)
• 🌱 machine translation (text-only and multimodal machine translation)
• 👯 multimodal learning (pretraining technology and reasoning)
• 🌱 large language modeling (infra, multilingual pretrain and efficient universal sft)

I am passionate about specializing in algorithms and fit them into practical applications.

• 📫 2023-09 - now : working on Foundational LLM Team, Alibaba Inc., towards the universal intelligence of LLM, especially on dialogue and searching.
• 📫 2022-04 - 2023-09: worked on ByteDance AI Lab in the fields of multimodal/multilingual machine translation and multilingual LLM.
• 📫 2021-07 - 2021-11: conducted research on semi-parametric MT as a NLP Research intern on Alibaba Damo Academy (One conference paper published).
• 📫 2020-11 - 2021-02: participated in early NLP Migration Project on HUAWEI Ascend, our work was reported as a markable practice [wiki].
• 📫 2020-05 - 2020-11: conducted research on translation quality estimation in corporation with OPPO Research (One paper under review).
• 📫 2020-04 - 2020-09: conducted research on vison & language multimodal machine translation (One conference paper published).
• 🤔 2019-09 - 2020-05: joined in TJUNLP lab and conducted research on vision & language commensense reasoning, finally stopped for the lack of computational resources.
• 👯 2018-03 - 2019-09: joined in Optimization Machine Learning Team and studied Fuzzy Clustering Theory (major) and Mainfold Learning (secondary) (One journal paper published and another two journal papers collaborated).
• 👯 2016-11 - 2018-09: joined the Drone Team in charge of the compute vision algorithm, won the second place in International Aerial Robotics Competition.

• Efficient Cluster-Based k-Nearest-Neighbor Machine Translation. ACL. 2022.
• AdaST: Dynamically Adapting Encoder States in the Decoder for End-to-End Speech-to-Text Translation. ACL Findings. 2021.
• Efficient Object-Level Visual Context Modeling for Multimodal Machine Translation: Masking Irrelevant Objects Helps Grounding. AAAI. 2021.
• A Novel Fuzzy c-Means Clustering Algorithm Using Adaptive Norm. International Journal of Fuzzy Sytems. 2019.