Huadong Feng

Huadong Feng

Auto-generated profile, based on publicly available information

Huadong Feng

Department of Computer Science and Engineering, The University of Texas at Arlington, Arlington, TX, USA

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 August 2020
Verified
Conference Paper
Closed access
JCJaganmohan Chandrasekaran
Jaganmohan Chandrasekaran
HFHuadong Feng
Huadong Feng
YLYu Lei
Yu Lei
RKRaghu Kacker
Raghu Kacker
DKD. Richard Kuhn
D. Richard Kuhn
Many machine learning algorithms examine large amounts of data to discover insights from hidden patterns. Testing these algorithms can be expensive and time-consuming. There is a need to speed up the testing process, especially in an agile development process, where testing is frequently performed. One approach is to replace big datasets with smaller datasets produced by random sampling. In this paper, we report a set of experiments that are designed to evaluate the effectiveness of using reduced datasets produced by random sampling for testing machine learning algorithms. In our experiments, we use as subject programs four supervised learning algorithms from the Waikato Environment for Knowledge Analysis (WEKA). We identify five datasets from Kaggle.com to run with the four learning algorithms. For each dataset, we generate reduced datasets of different sizes using two random sampling strategies, i.e., pure random and stratified random sampling. We execute our subject programs with the original and the reduced datasets, and measure test effectiveness using branch and mutation coverage. Our results indicate that in most cases, reduced datasets of even very small sizes can achieve the same or similar coverage achieved by the original dataset. Furthermore, our results indicate that reduced datasets produced by the two sample strategies do not differ significantly, and branch coverage correlates with mutation coverage.

7 March 2023 in Digital Threats: Research and Practice
Verified
Journal Article
Open access
MRMengfei Ren
Mengfei Ren
XRXiaolei Ren
Xiaolei Ren
HFHuadong Feng
Huadong Feng
JMJiang Ming
Jiang Ming
YLYu Lei
Yu Lei
Zigbee is widely adopted as a resource-efficient wireless protocol in the IoT network. IoT devices from manufacturers have recently been affected due to major vulnerabilities in Zigbee protocol implementations. Security testing of Zigbee protocol implementations is becoming increasingly important. However, applying existing vulnerability detection techniques such as fuzzing to the Zigbee protocol is not a simple task. Dealing with low-level hardware events still remains a big challenge. For the Zigbee protocol, which communicates over a radio channel, many existing protocol fuzzing tools lack a sufficient execution environment. To narrow the gap, we designed Z-Fuzzer , a device-agnostic fuzzing tool for detecting security flaws in Zigbee protocol implementations. To simulate Zigbee protocol execution, Z-Fuzzer leverages a commercial embedded device simulator with pre-defined peripherals and hardware interrupt setups to interact with the fuzzing engine. Z-Fuzzer generates more high-quality test cases with code-coverage heuristics. We compare Z-Fuzzer with advanced protocol fuzzing tools, BooFuzz and Peach fuzzer, on top of Z-Fuzzer’s simulation platform. Our findings suggest that Z-Fuzzer can achieve greater code coverage in Z-Stack, a widely used Zigbee protocol implementation. Compared to BooFuzz and Peach, Z-Fuzzer found more vulnerabilities with fewer test cases. Three of them have been assigned CVE IDs with high CVSS scores (7.5~8.2).

7 February 2025 in Distributed Ledger Technologies: Research and Practice
Verified
Journal Article
Open access
QWQiping Wei
Qiping Wei
FSFadul Sikder
Fadul Sikder
HFHuadong Feng
Huadong Feng
YLYu Lei
Yu Lei
RKRaghu Kacker
Raghu Kacker
RKRichard Kuhn
Richard Kuhn
Symbolic execution of smart contracts suffers from sequence explosion. Some existing tools limit the sequence length, thus being unable to adequately evaluate some functions. In this article, we propose a symbolic execution approach without limiting the sequence length. In our approach, the symbolic execution process is a two-phase model that maximizes code coverage while reducing the number of sequences to be executed. The first phase executes all sequences up to a length limit to identify the not-fully covered functions, while the second attempts to cover these functions according to state evaluation and a function graph structure. We have developed a tool called SmartExecutor and conducted an experimental evaluation on the SGUARD dataset. The experimental results indicate that compared with state-of-the-art tools, SmartExecutor achieves higher code coverage with less time. It also detects more vulnerabilities than Mythril, a state-of-the-art symbolic execution tool.