Chenglong Fu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chenglong Fu has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Xiaojiang Du
Stevens Institute of Technology, Hoboken, New Jersey, USA
6 shared publications - Huan Bui
The University of North Carolina at Charlotte
3 shared publications - Bo Luo
The University of Kansas, Lawrence, KS, USA
2 shared publications - Fei Zuo
University of Central Oklahoma
2 shared publications - Xuening Xu
Department of Electrical and Computer Engineering, Stevens Institute of Technology, Hoboken, NJ, USA
2 shared publications - Harper Lienerth
Department of Mathematics and Computer Science, Albion College, Albion, Michigan, USA
1 shared publication
Latest Publications
- Open Access
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 3310-3324
Jailbreaking via Activation Dilution: Neuron-Guided Adaptive Decomposition of Language Models
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 3310-3324
Open AccessJailbreaking via Activation Dilution: Neuron-Guided Adaptive Decomposition of Language Models
- 8 November 2026
- Published by Springer Nature
- p. 86-105
L2R: Verification-Guided Repair of LLM-Generated Smart-Home Automation Rules
- 8 November 2026
- Published by Springer Nature
- p. 86-105
L2R: Verification-Guided Repair of LLM-Generated Smart-Home Automation Rules
- 8 November 2026
- Published by Springer Nature
- p. 285-304
Does Efficiency Leak? Auditing the Privacy of Compressed and Stale Updates in Federated Intrusion Detection for IoT
- 8 November 2026
- Published by Springer Nature
- p. 285-304
Does Efficiency Leak? Auditing the Privacy of Compressed and Stale Updates in Federated Intrusion Detection for IoT
- 30 October 2026
- Published by Springer Nature
- p. 477-507
UDPU: Unlabeled Data Purification and Unlearning for Backdoor Defense
- 30 October 2026
- Published by Springer Nature
- p. 477-507
UDPU: Unlabeled Data Purification and Unlearning for Backdoor Defense
- 29 October 2026
- Published by Springer Nature
- p. 422-448
Effective Handling of Erroneous Executions of Automation Rules on Multi-platform Smart Home IoT Systems
- 29 October 2026
- Published by Springer Nature
- p. 422-448
Effective Handling of Erroneous Executions of Automation Rules on Multi-platform Smart Home IoT Systems
- Open Access
- 3 September 2026
- Published by Elsevier in SSRN Electronic Journal
Research Should Outlive the Agent: RKA as Researcher-controlled Project Memory for Agentic Research
Agentic research systems can now search literature, write and run code, run experiments, and draft manuscripts. They work, and that is the problem. A capable agent produces more plausible material in an afternoon than a researcher can follow in a week, all written in the same fluent register whether it records a measurement, a suggestion, or a decision. The bottleneck on a research project has moved from producing work to judging it. We present RKA Core, a local-first, model-neutral project record built for that problem. RKA is not a research agent. It separates what happened from what someone thinks it means from what the project has accepted, and hands the researcher assembled material at the moments a decision is due: options with their tradeoffs and reversibility stated, bounded checkpoints when an agent needs input, and the evidence behind any position, current or historical. Agents record activity as they work; the researcher frames, chooses, resolves, and revises. In a reconstruction study over one project history, independent clients answering from the structured record recovered every reference fact. The same clients working from searchable traces or a model-written summary recovered roughly 40%, and the summary condition was the most confident of the three and the only one to fabricate a detail. The paper explains the burden, shows a working day with the tool, states the design behind it, reports what we measured, and says what remains open.
- 3 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessResearch Should Outlive the Agent: RKA as Researcher-controlled Project Memory for Agentic Research
Agentic research systems can now search literature, write and run code, run experiments, and draft manuscripts. They work, and that is the problem. A capable agent produces more plausible material in an afternoon than a researcher can follow in a week, all written in the same fluent register whether it records a measurement, a suggestion, or a decision. The bottleneck on a research project has moved from producing work to judging it. We present RKA Core, a local-first, model-neutral project record built for that problem. RKA is not a research agent. It separates what happened from what someone thinks it means from what the project has accepted, and hands the researcher assembled material at the moments a decision is due: options with their tradeoffs and reversibility stated, bounded checkpoints when an agent needs input, and the evidence behind any position, current or historical. Agents record activity as they work; the researcher frames, chooses, resolves, and revises. In a reconstruction study over one project history, independent clients answering from the structured record recovered every reference fact. The same clients working from searchable traces or a model-written summary recovered roughly 40%, and the summary condition was the most confident of the three and the only one to fabricate a detail. The paper explains the burden, shows a working day with the tool, states the design behind it, reports what we measured, and says what remains open.
- 2 September 2026
- Published by Springer Nature
- p. 41-61
TAPAssure: Safe Automation in Smart Homes with LLMs and Formal Verification
- 2 September 2026
- Published by Springer Nature
- p. 41-61
TAPAssure: Safe Automation in Smart Homes with LLMs and Formal Verification
- Open Access
- 22 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 100-109
LLM-GridEval: Measuring the Evaluation Validity Gap in Smart Grid Security with Adaptive LLM Attackers
- 22 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 100-109
Open AccessLLM-GridEval: Measuring the Evaluation Validity Gap in Smart Grid Security with Adaptive LLM Attackers
- Open Access
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1649-1663
Discovering and Exploiting IoT Device Hidden Attributes: A New Vulnerability in Smart Homes
- 19 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 1649-1663
Open AccessDiscovering and Exploiting IoT Device Hidden Attributes: A New Vulnerability in Smart Homes
- 27 June 2025
- Published by Springer Nature
- p. 158-184
ShadowConn: Breaking the Entanglement of Cross-Platform IoT Delegation in Multi-user Environments
- 27 June 2025
- Published by Springer Nature
- p. 158-184
ShadowConn: Breaking the Entanglement of Cross-Platform IoT Delegation in Multi-user Environments