Semantics Preserving Emoji Recommendation with Large Language Models

Semantics Preserving Emoji Recommendation with Large Language Models

Zhongyi Qiu, Kangyi Qiu, Hanjia Lyu, Wei Xiong, Jiebo Luo

Accepted for publication in IEEE BigData 2024

Contact

Zhongyi Qiu ([email protected]), Kangyi Qiu ([email protected]), Hanjia Lyu ([email protected]), Wei Xiong ([email protected]), Jiebo Luo ([email protected])

Environment Setup for Emoji Recommendation Project

Prerequisites

1. Ensure Conda is installed. Install Conda if not already available.
2. Ensure requirements.txt is in the same directory as setup_env.sh.

Steps to Set Up the Environment

1. Run the setup script:
   Use the following command to run the setup script:

   bash setup_env.sh

   Or:

   sh setup_env.sh

   This creates the emoji_recommendation environment, installs Python 3.10, and installs dependencies from requirements.txt.

2. Activate the environment:

   conda activate emoji_recommendation

Folder Descriptions

• Code:

  • The src folder contains all the Jupyter Notebook (.ipynb) files used for data preprocessing and running the models. These notebooks handle tasks such as preparing the dataset, running the emoji recommendation models, and evaluating the results.

• Dataset:

  • src/dataset: Contains the initial raw dataset used for preprocessing and analysis.
  • src/dataset_finalOutput: Stores the output results generated by six large language models (LLMs).
  • src/human_eval: Includes a small subset of manually labeled data used for evaluating the model predictions.

Citation

@article{qiu2024semantics,
  title={Semantics Preserving Emoji Recommendation with Large Language Models},
  author={Qiu, Zhongyi and Qiu, Kangyi and Lyu, Hanjia and Xiong, Wei and Luo, Jiebo},
  journal={arXiv preprint arXiv:2409.10760},
  year={2024}
}

Related Work

[ICWSM 2024] Human vs. LMMs: Exploring the Discrepancy in Emoji Interpretation and Usage in Digital Communication

[ACM TIST 2024] GPT-4V(ision) as A Social Media Analysis Engine

[ICPR 2024] A Benchmark and Chain-of-Thought Prompting Strategy for Large Multimodal Models with Multiple Image Inputs

[ICME 2024] Chain-of-Thought Prompting for Demographic Inference with Large Multimodal Models

[WWW 2024] Unifying Local and Global Knowledge: Empowering Large Language Models as Political Experts with Knowledge Graphs

[ACL 2024] SoMeLVLM: A Large Vision Language Model for Social Media Processing

[COLING 2025] Evolver: Chain-of-Evolution Prompting to Boost Large Multimodal Models for Hateful Meme Detection