I am a PhD student at the University of Exeter. My research interests include natural language processing, computer vision, bioinformatics (RNA modeling), etc. Please find the details of my study in the following sections. Find me on GitHub or Google Scholar.

• Textual Adversarial Attack/Defense
• Sentiment (Coherency) Analysis
• bio-RNA Sequence pretraining & modeling
• Code language modeling
• Diffusion models

EACL 2024 Findings H Yang, K Li. Improving Implicit Sentiment Learning via Local Sentiment Aggregation

ACL 2023 Findings H Yang, K Li. Boosting Text Augmentation via Hybrid Instance Filtering Framework

EMNLP 2023 Findings H Yang, K Li. InstOptima: Evolutionary Multi-objective Instruction Optimization via Large Language Model-based Instruction Operators

CIKM 2023 H Yang, K Li. PyABSA: A Modularized Framework for Reproducible Aspect-based Sentiment Analysis

IEEE Trans. Softw. Eng. K Li, H Yang, W Visser. Evolutionary Multi-Task Injection Testing on Web Application Firewalls

J Supercomput B Zeng, H Yang, S Liu, M Xu. Learning for target-dependent sentiment based on local context-aware embedding

Neurocomputing H Yang, B Zeng, JH Yang, Y Song, R Xu. A multi-task learning model for Chinese-oriented aspect polarity classification and aspect term extraction

H Yang, K Li. Reactive Perturbation Defocusing for Textual Adversarial Defense (Preprint 2023)

B Zeng, H Yang, R Xu, W Zhou, X Han. LCF: A local context focus mechanism for aspect-based sentiment classification (Applied Sciences 2019).

K Li is my PhD's supervisor. B Zeng was my master's supervisor.

Neurocomputing M Xu, B Zeng, H Yang, J Chi, J Chen, H Liu. Combining dynamic local context focus and dependency cluster attention for aspect-level sentiment classification

• Aspect-based sentiment analysis (ABSA)
• Super-resolution anime diffusion

• ABSA Model:yangheng/deberta-v3-base-absa-v1.1
• ABSA Model:yangheng/deberta-v3-large-absa-v1.1

• PyABSA: State-of-the-art models and simplified interface for ABSA
• FindFile: A simple tool to help find your file wherever it locates
• Metric-Visualizer: Help you to easily and quickly make academic plots, e.g., box and trajectory plots
• BoostAug: it can achieve at most 5% absolute improvement based on full-scale datasets for most models.
• InstOptima: Evolutionary Multi-objective Instruction Optimization via Large Language Model-based Instruction Operators.
• Adversarial defense: This work propose an effective adversarial defense method for textual adversarial defense.
• ABSA DPT: This is a web-based ABSA dataset annotation tool that can help you easily make an ABSA dataset and train it using PyABSA.
• [More works are going to be published]