David Lo

Making privacy policies easier to read and understand for mobile app users.

A study comparing the quality of human and AI responses in programming.

This study compares PEFT and ICL in improving code generation using LLMs.

A novel framework improves the effectiveness of adversarial attacks on code models.

Examining current tools for method naming consistency reveals significant shortcomings.

A technique to reduce code model size while maintaining performance.

A new model improves how we work with code using intermediate representations.

A method to improve understanding and security of Graph Neural Networks.

SeePrivacy aims to simplify how users interact with privacy policies.

FuseFL enhances fault localization by combining various data sources for clearer explanations.

A novel system for assessing software vulnerabilities at multiple levels.

A look into large language models for coding and their ecosystem.

This study assesses the effectiveness of LLMs in creating unit tests.

New models help developers, but backdoor attacks pose serious security risks.

Exploring adaptive systems for managing runtime errors using Large Language Models.

A new approach to enhance tagging accuracy using pre-trained models.

Examining security concerns in user feedback mechanisms for code generation tools.

A new method to enhance code generation in LLM4Code models.

A detailed look at breaking changes and their impact on developers.

This study compares different LLMs for finding faults and detecting vulnerabilities in software.

Learn how Parameter-Efficient Fine-Tuning improves code smell detection with LLMs.

ACECode optimizes AI-generated code for efficiency and correctness.