Latest preprints

This research evaluates the extent to which modern AI frameworks, specifically TensorFlow and scikit-learn, adhere to the SOLID design principles—Single Responsibility, Open/Closed, Liskov Substitution, Interface Segregation, and Dependency Inversion. Analyzing the frameworks' architectural documentation and design philosophies, this research investigates architectural trade-offs when balancing software engineering best practices with AI-specific needs. I examined each framework's documentation, source code, and architectural components to evaluate their adherence to these principles. The results show that both frameworks adopt certain aspects of SOLID design principles but make intentional trade-offs to address performance, scalability, and the experimental nature of AI development. TensorFlow focuses on performance and scalability, sometimes sacrificing strict adherence to principles like Single Responsibility and Interface Segregation. While scikit-learn’s design philosophy aligns more closely with SOLID principles through consistent interfaces and composition principles, it sticks closer to SOLID guidelines but with occasional deviations for performance optimizations and scalability. This research discovered that applying SOLID principles in AI frameworks depends on context, as performance, scalability, and flexibility often require deviations from traditional software engineering principles. This research contributes to understanding how domain-specific constraints influence architectural decisions in modern AI frameworks and how these frameworks strategically adapted design choices to effectively balance these contradicting requirements.

AI for software engineering has made remarkable progress recently, becoming a notable success within generative AI. Despite this, there are still many challenges that need to be addressed before automated software engineering reaches its full potential. It should be possible to reach high levels of automation where humans can focus on the critical decisions of what to build and how to balance difficult tradeoffs while most routine development effort is automated away. Reaching this level of automation will require substantial research and engineering efforts across academia and industry. In this paper, we aim to discuss progress towards this in a threefold manner. First, we provide a structured taxonomy of concrete tasks in AI for software engineering, emphasizing the many other tasks in software engineering beyond code generation and completion. Second, we outline several key bottlenecks that limit current approaches. Finally, we provide an opinionated list of promising research directions toward making progress on these bottlenecks, hoping to inspire future research in this rapidly maturing field.

Large Language Models (LLMs) are of great interest in vulnerability detection and repair. The effectiveness of these models hinges on the quality of the datasets used for both training and evaluation. Our investigation reveals that a number of studies featured in prominent software engineering conferences have employed datasets that are plagued by high duplication rates, questionable label accuracy, and incomplete samples. Using these datasets for experimentation will yield incorrect results that are significantly different from actual expected behavior. For example, the state-of-the-art VulRepair Model, which is reported to have 44% accuracy, on average yielded 9% accuracy when test-set duplicates were removed from its training set and 13% accuracy when training-set duplicates were removed from its test set. In an effort to tackle these data quality concerns, we have retrained models from several papers without duplicates and conducted an accuracy assessment of labels for the top ten most hazardous Common Weakness Enumerations (CWEs). Our findings indicate that 56% of the samples had incorrect labels and 44% comprised incomplete samples—only 31% were both accurate and complete. Finally, we employ transfer learning using a large deduplicated bug-fix corpus to show that these models can exhibit better performance if given larger amounts of high-quality pre-training data, leading us to conclude that while previous studies have over-estimated performance due to poor dataset quality, this does not demonstrate that better performance is not possible.

Ingenuity is an autonomous Cyber-Pysical System (CPS) that has successfully completed more than 70 flights over Mars between 2021 and 2024. Ensuring the safety of its mission is paramount, as any failure could result in catastrophic economic damage and significant financial losses. Dataflow Models of Computation and Communication (DF MoCCs) serve as a formal framework for specifying and analyzing the timing behavior of such CPSs. In particular, the Real-time Mode-aware Dataflow (RMDF) model is highly suitable to specify and analyze real-time and mode-dependent CPSs like Ingenuity. This paper showcases the application of RMDF for the specification and analysis of Ingenuity. We propose a dataflow specification of Ingenuity, analyze its timing behavior, and provide a feasibility test. Finally, we proposed a plausible explanation of the timing anomaly that occurred during the sixth flight of Ingenuity.

Compilers are complex, and significant effort has been expended on testing them. Techniques such as random program generation and differential testing have proved highly effective and have uncovered thousands of bugs in production compilers. The majority of effort has been expended on validating that a compiler produces correct code for a given input, while less attention has been paid to ensuring that the compiler produces performant code.In this work we adapt differential testing to the task of identifying missed optimization opportunities in compilers. We develop a novel testing approach which combines large language models (LLMs) with a series of differential testing strategies and use them to find missing code size optimizations in C / C++ compilers.The advantage of our approach is its simplicity. We offload the complex task of generating random code to an off-the-shelf LLM, and use heuristics and analyses to identify anomalous compiler behavior. Our approach requires fewer than 150 lines of code to implement. This simplicity makes it extensible. By simply changing the target compiler and initial LLM prompt we port the approach from C / C++ to Rust and Swift, finding bugs in both. To date we have reported 24 confirmed bugs in production compilers, and conclude that LLM-assisted testing is a promising avenue for detecting optimization bugs in real world compilers.

Software vulnerabilities can result in catastrophic cyberattacks that increasingly threaten business operations. Consequently, ensuring the safety of software systems has become a paramount concern for both private and public sectors. Recent literature has witnessed increasing exploration of learning-based approaches for software vulnerability detection. However, a key limitation of these techniques is their primary focus on a single programming language, such as C/C++, which poses constraints considering the polyglot nature of modern software projects. Further, there appears to be an oversight in harnessing the synergies of vulnerability knowledge across varied languages, potentially underutilizing the full capabilities of these methods. To address the aforementioned issues, we introduce MVD – an innovative multi-lingual vulnerability detection framework. This framework acquires the ability to detect vulnerabilities across multiple languages by concurrently learning from vulnerability data of various languages, which are curated by our specialized pipeline. We also incorporate incremental learning to enable the detection capability of MVD to be extended to new languages, thus augmenting its practical utility. Extensive experiments on our curated dataset of more than 11K real-world multi-lingual vulnerabilities substantiate that our framework significantly surpasses state-of-the-art methods in multi-lingual vulnerability detection by 83.7% to 193.6% in PR-AUC. The results also demonstrate that MVD detects vulnerabilities well for new languages without compromising the detection performance of previously trained languages, even when training data for the older languages is unavailable. Overall, our findings motivate and pave the way for the prediction of multi-lingual vulnerabilities in modern software systems.

The paper presents a systematic methodology for analyzing software developer productivity by refining contribution rate metrics to distinguish meaningful development efforts from anomalies. Using the Mean-High Model Contribution Rate (mhMCR) method, the research introduces a statistical framework that focuses on continuous contributions, mitigating distortions caused by tool-assisted refactoring, delayed commits, or automated changes. The methodology integrates clustering techniques, commit time deltas, and contribution sizes to isolate natural, logical work patterns and supports the accurate imputation of effort for contributions outside these patterns. Through empirical validation across multiple commercial repositories, the mhMCR method demonstrates enhanced precision in productivity measurement in identifying sustained developer activity. The findings provide actionable insights for optimizing team performance and workflow management in modern software engineering practices.

In the advanced technology nodes, the integrated design rule checker (DRC) is often utilized in place and route tools for fast optimization loops for power-performance-area. Implementing integrated DRC checkers to meet the standard of commercial DRC tools demands extensive human expertise to interpret foundry specifications, analyze layouts, and debug code iteratively. However, this labor-intensive process, requiring to be repeated by every update of technology nodes, prolongs the turnaround time of designing circuits.In this paper, we present DRC-Coder, a multi-agent framework with vision capabilities for automated DRC code generation. By incorporating vision language models and large language models (LLM), DRC-Coder can effectively process textual, visual, and layout information to perform rule interpretation and coding by two specialized LLMs. We also design an auto-evaluation function for LLMs to enable DRC code debugging. Experimental results show that targeting on a sub-3nm technology node for a state-of-the-art standard cell layout tool, DRC-Coder achieves perfect F1 score 1.000 in generating DRC codes for meeting the standard of a commercial DRC tool, highly outperforming standard prompting techniques (F1=0.631). DRC-Coder can generate code for each design rule within four minutes on average, which significantly accelerates technology advancement and reduces engineering costs.