Hi there!

I am a PhD student at the University of Exeter, specializing in genomics, bioinformatics, and Large Language Models (LLMs). My research includes biological sequence modeling, sentiment analysis, adversarial attacks, and open-source tool development. I’ve contributed to the field with OmniGenomeBench and PyABSA, and have published in leading conferences like ACL, EMNLP, CIKM, EACL, IEEE TSE, etc.

I actively contribute to platforms like GitHub and Hugging Face, sharing tools like FindFile, MetricVisualizer, BoostAug. I’m committed to advancing AI and NLP while making these technologies accessible through open-source contributions.

I expect to graduate in 2025 and am open to new opportunities in academia and industry.

• OmniGenomeBench: Automating Large-scale in-silico Benchmarking for Genomic Foundation Models
  Heng Yang, Jack Cole, Ke Li | Arxiv 2024

• OmniGenome: Aligning RNA Sequences with Secondary Structures in Genomic Foundation Models
  Heng Yang, Ke Li | Arxiv 2024

• The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples
  Heng Yang, Ke Li | EMNLP 2024

• MP-RNA: Unleashing Multi-species RNA Foundation Model via Calibrated Secondary Structure Prediction
  Heng Yang, Ke Li | EMNLP 2024

• Modeling Aspect Sentiment Coherency via Local Sentiment Aggregation
  Heng Yang, Ke Li | EACL 2024

• PyABSA: A Modularized Framework for Reproducible Aspect-based Sentiment Analysis
  Heng Yang, Cheng Zhang, Ke Li | CIKM 2023

• InstOptima: Evolutionary Multi-objective Instruction Optimization via Large Language Model-based Instruction Operators
  Heng Yang, Ke Li | EMNLP 2023

• Boosting Text Augmentation via Hybrid Instance Filtering Framework
  Heng Yang, Ke Li | ACL 2023

• DaNuoYi: Evolutionary Multi-Task Injection Testing on Web Application Firewalls
  Ke Li, Heng Yang | IEEE TSE, 2023

• A Multi-task Learning Model for Chinese-oriented Aspect Polarity Classification and Aspect Term Extraction
  Heng Yang, Biqing Zeng, et al. | Neurocomputing, 2021

A large-scale in-silico benchmarking framework for genomic foundation models (GFMs). It addresses the lack of standardized tools for evaluating GFMs, automating the benchmarking process for diverse models. Features a public leaderboard for tracking performance across models.

A modularized framework for Aspect-Based Sentiment Analysis (ABSA). PyABSA simplifies sentiment analysis tasks with pre-trained models and datasets for research and production environments.

• DeBERTa-v3 Base ABSA
  A model for aspect-based sentiment analysis (ABSA), trained with over 30k samples for tasks like sentiment classification.

• PlantRNA-FM
  An interpretable RNA foundation model for exploring functional RNA motifs in plants. Pre-trained on data from over 1,124 plant species.

• OmniGenome
  A genomic model aimed at biological sequence modeling, part of the OmniGenome project.

Profile Views: