Yibo Chen

Staff Scientist

Washington University in St. Louis
Country flagUnited States

Research Interests

ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

Articles (5)

Large language models (LLMs) and prompt-based learning as transformative tools have been widely used across both academic and industrial domains. This paper reviews the current landscape of LLM applications, examining their principles, methodologies, strengths, and limitations. Beginning with an overview of the evolution from traditional feature engineering to prompt-based approaches, it highlights the applications of LLMs in diverse fields, including bioinformatics, materials science, and drug discovery. The review further examines the technical intricacies of prompt-based learning, contrasting hard and soft prompt techniques and their respective contributions to optimizing model performance. Real-world implementations are analyzed, with a focus on applications in bioinformatics and financial technology, alongside a discussion of pressing challenges related to data security and privacy. This paper also investigates the economic value generated by LLMs and their potential societal and workforce impacts, such as industry disruption and new job creation. The anticipated trajectory of LLMs and their broader societal implications are discussed, emphasizing the need for continued research and policy to guide its responsible development.

Year:

2025

Year:

2025

Single-cell large language models (scLLMs) capture essential biological insights from vast single-cell atlases but struggle in out-of-context applications, where zero-shot predictions can be unreliable. To address this, we introduce a single-cell parameter-efficient fine-tuning (scPEFT) framework that integrates learnable, low-dimensional adapters into scLLMs. By freezing the backbone model and updating only the adapter parameters, scPEFT efficiently adapts to specific tasks using limited custom data. This approach mitigates catastrophic forgetting, reduces parameter tuning by over 96%, and decreases GPU memory usage by more than half, significantly enhancing scLLMs’s accessibility for resource-constrained researchers. Validated across diverse datasets, scPEFT outperformed zero-shot models and traditional fine-tuning in disease-specific, cross-species, and under-characterized cell population tasks. Its attention-mechanism analysis identified COVID-related genes associated with specific cell states and uncovered unique blood cell subpopulations, demonstrating scPEFT’s capacity for condition-specific interpretations. These findings position scPEFT as an efficient solution for improving scLLMs’ utilities in general single-cell analyses.

Year:

2025

Year:

2024

ChatGPT has demonstrated its potential as a surrogate knowledge graph. Trained on extensive data sources, including open-access publications, peer-reviewed research articles, and biomedical websites, ChatGPT extracted information on gene relationships and biological pathways so that it can be used to predict them. However, a major challenge is model hallucination, that is, high false positive rates. To assess and address this challenge, we systematically evaluated ChatGPT’s capacity for predicting gene relationships using GPT-3.5-turbo, GPT-4, and GPT-4o. Benchmarking against the KEGG Pathway Database as the ground truth, we experimented with diverse prompting strategies, targeting gene relationships of activation, inhibition, and phosphorylation. We introduced an innovative iterative prompt refinement technique. By assessing prompt efficacy using metrics such as F-1 score, precision, and recall, GPT-4 suggested improved prompts. A refined prompt, which combines a specialized role with explanatory text, significantly enhanced the performance. Going beyond pairwise gene relationships, we also deciphered complex gene interplays, such as gene interaction chains and pathways pertinent to diseases such as non-small cell lung cancer. Direct prompts showed limited success, but “least-to-most” prompting exhibited significant potentials for such network constructions. The methods in this study may be used for other bioinformatics prediction problems.

Year:

2024

Collaborators (5)

Mihail Popescu

Associate Professor

University of Missouri Columbia

UNITED STATES
View details

Gangqing Hu

Associate Professor and Director of the Bioinformatics Core

West Virginia University

UNITED STATES
View details

Anjun Ma

Assistant Professor

Ohio State University

UNITED STATES
View details

Qin Ma

Vice Chair for Research

Ohio State University

UNITED STATES
View details

Dong Xu

Professor

University of South Florida

UNITED STATES
View details