Northeast University(NEU), ShenYang, China
2020.09 - 2024.06, Undergraduate, Software Engineering

Sun Yat-sen University(SYSU), Kwangtung, China
2024.09 - XXXX.XX, Master, Computer Science

NEU Data Mining Group

2022.10 ~ 2022.04

Field of Research:

• Multimodal Sentiment Analysis
• Large Empathic Dialogue Model

Participating Work:

• PICA: Unleashing The Emotional Power of Large Language Model

FDU Data Intelligence and Social Computing Lab

2023.02 - Now

Field of Research:

• Large Vision Language Model Benchmark
• Large Empathic Dialogue Model Interpretability Research

Participating Work:

• ReForm-Eval: Evaluating large vision language models via unified re-formulation of task-oriented benckmarks
• Multimodal large model probe based on LLaVA(Ongoing)

SYSU Human Cyber Physical Intelligence Integration Lab

2023.10 - Now (Maybe I am in)

Field of Research:

• Long Video Genearation based on LaVie(Ongoing)

Participating Work:

Huggingface

• Fine-tune BERT model on SST2 data set to implement text classification
• Fine-tune the ViT model on the Food101 dataset for image classification
• Fine-tune the Wav2Vec model on the MInDS-14 dataset for audio classification

Kaggle

• Sentiment Analysis on Movie Review (link)
  

  Fine-grained sentiment classification of film review data

• Natrual Language Processing with Disaster Tweets (link)
  

  Dichotomies predict whether a tweet is about a real disaster

• Google AI4Code - Understand Code in Python Notebooks (link)
  

  Sort markdown cells according to the code cell order in the json file of the Jupyter notebook

Natural Language Processing (code)

• Spelling Correction
  

  Based on Bayesian probability formula and Markov chain

• Word Embedding
  

  Word vector calculation based on Skip-Gram or CBOW

• Fine-Grained Text Classification
  

  Linear regression classification, dataset, gradient descent, etc. based on numpy

• Fine-Grained Text Classification
  

  Use the pytorch framework to implement classification using TextCNN or RNN series models

Cross-Visual Language Multimodal

• Image and text retrieval system (code)
  

  Based on the dual-tower architecture , use GRU model, ResNet model, and other transformer architecture pre-training models to try to improve

• ReForm-Eval (code)
  

  Evaluating large vision language models via unified re-formulation of task-oriented benckmarks. Participated during internship in FDU-DISC laboratory

Front-End Development

• Hanoi Tower (code)
  
• Large Screen Statistics System (code)