Miles Q. Li (李琦)

Miles Q. Li (李琦)

Ph.D., AI researcher, AI Agent System Developer

Website Hit Counters

Publications

Feedback-guided Post-training of Large Language Models

Published in Communications AI & Computing (Springer Nature), accepted, in press, 2026

A Review that organizes feedback-guided post-training methods for LLMs (RLHF, RLVR, DPO-style and related algorithms) under a single taxonomy and shared notation.

Recommended citation: Miles Q. Li, Benjamin C. M. Fung, Ebrahim Bagheri, Shih-Chia Huang, and Ching-Chun Huang. Feedback-guided Post-training of Large Language Models. Communications AI & Computing (2026), in press.

GPT-based Self-supervised Anomaly Detection in Command Lines

Published in Journal of Computer Virology and Hacking Techniques, 2026

This paper is about our GPT-based self-supervised anomaly detection system for command lines

Recommended citation: Miles Q. Li, Julien Keutchayan, François Charest, and Benjamin C.M. Fung. GPT-based Self-supervised Anomaly Detection in Command Lines. Journal of Computer Virology and Hacking Techniques, Springer, 2026.

A Benchmark for Evaluating Outcome-Driven Constraint Violations in Autonomous AI Agents

Published in arXiv preprint / Under review for IEEE Transactions on Artificial Intelligence, 2025

This paper presents ODCV-Bench, a safety benchmark for evaluating constraint violations in autonomous AI agents. Reached #1 on Hacker News in February 2026 (544 points, 366 comments): discussion · ranking history.

Recommended citation: Miles Q. Li, Benjamin Fung, Martin Weiss, Pulei Xiong, Khalil Al-Hussaeni, and Claude Fachkha. A Benchmark for Evaluating Outcome-Driven Constraint Violations in Autonomous AI Agents. arXiv preprint arXiv:2512.20798 (2025). https://arxiv.org/abs/2512.20798

Security Concerns for Large Language Models: A Survey

Published in Journal of Information Security and Applications, 2025

This paper is about our study on the security concerns with LLMs

Recommended citation: Miles Q. Li, and Benjamin CM Fung. Security Concerns for Large Language Models: A Survey. Journal of Information Security and Applications 95 (2025): 104284. https://www.sciencedirect.com/science/article/pii/S2214212625003217?casa_token=8Ce8QlKHMEoAAAAA:Dy_eO6f0zDbNjuXcwnPBnT9ezs0QQu8Ne_sn1DThh55aw4u-QP4OL0PbOIWzlL_ydi8uhlsP4w

VDGraph2Vec: Vulnerability Detection in Assembly Code Using Message Passing Neural Networks

Published in ICMLA 2022, 2022

This paper proposes an automated deep learning method to generate representations of assembly code for vulnerability detection.

Recommended citation: Ashita Diwan, Miles Q. Li, and Benjamin CM Fung. "VDGraph2Vec: Vulnerability Detection in Assembly Code Using Message Passing Neural Networks." In 2022 21st IEEE International Conference on Machine Learning and Applications (ICMLA), pp. 1039-1046. IEEE, 2022. https://ieeexplore.ieee.org/document/10069134/

DyAdvDefender: An instance-based online machine learning model for perturbation-trial-based black-box adversarial defense

Published in Journal of Information Sciences, 2022

This paper is about our novel black-box adversarial defense method

Recommended citation: Li, Miles Q., Benjamin CM Fung, and Philippe Charland. "DyAdvDefender: An instance-based online machine learning model for perturbation-trial-based black-box adversarial defense." Information Sciences (2022). https://www.sciencedirect.com/science/article/pii/S0020025522003747?casa_token=p5N50hWOf0oAAAAA:OoG3up9I8-W8kW1zutzK3zuzOZL1kpWspm_7h0YJZC_aowNcFvN97aUNwcWJvMX61QngMi4aNjy4

A Novel Neural Network-Based Malware Severity Classification System

Published in International Conference on Software Technologies (Springer), 2021

This paper proposes a neural network-based malware severity classification method.

Recommended citation: Miles Q. Li and Benjamin CM Fung. "A Novel Neural Network-Based Malware Severity Classification System." In International Conference on Software Technologies, pp. 218-232. Springer, 2021. https://link.springer.com/chapter/10.1007/978-3-031-11513-4_10