28 April 2025 in ACM Transactions on Software Engineering and Methodology
Predicting function names in stripped binaries, which requires succinctly summarizing semantics of binary code in natural languages, is a crucial but challenging task. Recently, many machine learning based solutions have been proposed. However, they have poor generalizability, i.e., fail to handle unseen binaries. To advance the state of the art, we present
large assembly language Model (
llasm
)
, a novel framework which fuses encoder-only and decoder-only LLMs for function name prediction. It refines encoder-only models to preserve more binary information and learn better binary representations. Then it adopts a novel architecture to project the encoding to the input space of a decoder-only natural language model, which enables it to have better capability of inferring general knowledge and better generalizability. We have evaluated llasm in the BinaryCorp and Debin datasets. llasm outperforms the state-of-the-art function name prediction tools by up to 19.9%, 40.7%, and 36.5% in precision, recall, and F1 score, with significantly better generalizability in unseen binaries. Our case studies further demonstrate the practical use cases of llasm in analyzing real-world malware, showing the usefulness of function name prediction.
12 March 2026 in ACM Transactions on Software Engineering and Methodology
File access vulnerabilities (FAVs) are one type of security weakness arising from adversary manipulations of file access inputs, posing significant threats to system integrity. Despite their prevalence, FAVs remain underexplored due to limited understanding, complex triggering scenarios, and stealthy and diverse manifestations; these challenges render current detection approaches incomplete and inaccurate. To this end, we conducted an in-depth empirical study across 204 file-related CVEs, uncovering the root cause and trigger mechanisms of FAVs. Based on these findings, we propose an exhaustive accessing model and a specialized threat model that define the adversary and attack surface for FAVs, enabling systematic attribution and analysis of file operations. Furthermore, we propose FAVDisco , a novel framework for discovering FAVs by mutating, triggering, and analyzing file operations. It employs a File Mutator to simulate diverse execution scenarios and an FAV Checker that integrates a model-based adversary controllable checker with pattern-based detection rules to identify FAVs. Implemented on Windows, FAVDisco achieves remarkable performance with 92.1% precision and 83.3% recall on the disclosed FAV detection task, outperforming state-of-the-art methods. Moreover, it uncovers 13 zero-day FAVs in 10 widely used services, with six assigned new CVEs and earning a reward of $29,000 from Microsoft Security Response Center.
22 June 2025 in Proceedings of the ACM on Software Engineering
Binary Function Retrieval (BFR) is crucial in reverse engineering for identifying specific functions in binary code, especially those associated with malicious behavior or vulnerabilities. Traditional BFR methods rely on heuristics, often lacking the efficiency and adaptability needed for large-scale or diverse binary analysis tasks. To address these challenges, we present BinQuery, a Natural Language-based BFR (NL-based BFR) framework that uses natural language queries to retrieve relevant binary functions with improved flexibility and precision. BinQuery introduces innovative techniques to bridge information gaps between binary code and natural language, achieves fine-grained alignment for enhanced retrieval accuracy, and leverages Large Language Models (LLMs) to refine queries and generate diverse descriptions. Our extensive experiments indicate that BinQuery surpasses current state-of-the-art methods, achieving a 42.55% increase in recall@1 and a 4× improvement in performance on comparable benchmarks.