Stefan Larson, Anish Mahendran, Joseph J. Peper, Christopher Clarke, Andrew Lee, Parker Hill, Jonathan K. Kummerfeld, Kevin Leach, Michael A. Laurenzano, Lingjia Tang, Jason Mars. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.

Mental health problems are highly prevalent and appear to be increasing in frequency severity among the college student population. The upsurge mobile wearable wireless technologies capable of intense, longitudinal tracking individuals, provide valuable opportunities examine temporal patterns dynamic interactions key variables mental research. In this paper, we present a feasibility study leveraging non-invasive sensing technology passively assess students' social anxiety, one most common...

With the rapid proliferation of malware attacks on Internet, understanding these malicious behaviors plays a critical role in crafting effective defense. Advanced analysis relies virtualization or emulation technology to run samples confined environment, and analyze activities by instrumenting code execution. However, virtual machines emulators inevitably create artifacts execution making approaches vulnerable detection subversion. In this paper, we present MALT, debugging framework that...

Software developers spend a great deal of time reading and understanding code that is poorly-documented, written by other developers, or developed using differing styles. During the past decade, researchers have investigated techniques for automatically documenting to improve comprehensibility. In particular, recent advances in deep learning led sophisticated summary generation convert functions methods simple English strings succinctly describe code's behavior. However, automatic...

Recent language models have demonstrated proficiency in summarizing source code. However, as many other domains of machine learning, code lack sufficient explainability --- informally, we a formulaic or intuitive understanding what and how learn from Explainability can be partially provided if, the to produce higher-quality summaries, they also align deeming same parts important those identified by human programmers. In this paper, report negative results our investigation summarization...

Virtual Machine Introspection (VMI) systems have been widely adopted for malware detection and analysis. VMI use hypervisor technology system introspection to expose malicious activity. However, recent can detect the presence of virtualization or corrupt state thus avoiding detection. We introduce SPECTRE, a hardware-assisted dependability framework that leverages System Management Mode (SMM) inspect system. Contrary VMI, our trusted code base is limited BIOS SMM implementations. SPECTRE...

Optimizing software performance through automated code refinement offers a promising avenue for enhancing execution speed and efficiency. Despite recent advancements in LLMs, significant gap remains their ability to perform in-depth program analysis. This study introduces AUTOPATCH, an in-context learning approach designed bridge this by enabling LLMs automatically generate optimized code. Inspired how programmers learn apply knowledge optimize software, AUTOPATCH incorporates three key...

Software engineering involves writing new code or editing existing code. Recent efforts have investigated the neural processes associated with reading and comprehending --- however, we lack a thorough understanding of human cognitive underlying writing. While prose been studied thoroughly, that same scrutiny has not applied to In this paper, leverage functional brain imaging investigate representations in comparison We present first study which participants wrote while undergoing magnetic...

Data structures permeate many aspects of software engineering, but their associated human cognitive processes are not thoroughly understood. We leverage medical imaging and insights from the psychological notion spatial ability to decode neural representations several fundamental data manipulations. In a study involving 76 participants, we examine list, array, tree, mental rotation tasks using both functional near-infrared spectroscopy (fNIRS) magnetic resonance (fMRI). find nuanced...

Code review is a critical step in modern software quality assurance, yet it vulnerable to human biases. Previous studies have clarified the extent of problem, particularly regarding biases against authors code,but no consensus understanding has emerged. Advances medical imaging are increasingly applied engineering, supporting grounded neurobiological explorations computing activities, including review, reading, and writing source code. In this paper, we present results controlled experiment...

A wide range of Arm endpoints leverage integrated and discrete GPUs to accelerate computation such as image processing numerical applications. However, in spite these important use cases, GPU security has yet be scrutinized by the community. By exploiting vulnerabilities kernel, attackers can directly access sensitive data used during computing, personally-identifiable computer vision tasks. Existing work Trusted Execution Environments (TEEs) address concerns on Intel-based platforms, while...

Confidential computing is an emerging technique that provides users and third-party developers with isolated transparent execution environment.To support this technique, Arm introduced the Computing Architecture (CCA), which creates multiple address spaces, known as realms, to ensure data confidentiality integrity in securitysensitive tasks.Arm recently proposed concept of confidential on GPU hardware, widely used generalpurpose, high-performance, artificial intelligence scenarios.However,...

In the evolving realm of natural language processing (NLP), generative AI models like ChatGPT are increasingly utilized across various applications. Among possible purposes, many people considering asking for relationship advice. However, lack in-depth examination ChatGPT's response quality could be concerning when it is used personal topics mental health issues and intimate problems. these topics, a piece misleading advice cause harmful repercussions. to people's growing interest in using...

With the growing ubiquity of uncrewed aerial vehicles (UAVs), mitigating emergent threats in such systems has become increasingly important. In this short paper, we discuss an indicative class UAVs and a potential attack scenario which benign UAV completing mission can be compromised by malicious attacker with antenna commodity computer open-source ground station software. We attest to relevance for both enterprise defense applications. describe system architecture resiliency trustworthiness...

According to a 2014 Spring American College Health Association Survey, almost 50% of college students reported feeling things were hopeless and that it was difficult function within the last 12 months. More than 80% overwhelmed exhausted by their responsibilities. This critical subpopulation Americans is facing significant levels mental health disorders, challenging colleges provide accessible high quality behavioral care. However, psychiatric disorders are frequently unrecognized in primary...

Understanding how developers carry out different computer science activities with objective measures can help to improve productivity and guide the use development of supporting tools in software engineering. In this article, we present two controlled experiments involving 112 students explore multiple computing (code comprehension, code review, data structure manipulations) using three including neuroimaging (functional near-infrared spectroscopy (fNIRS) functional magnetic resonance...

Neural code summarization leverages deep learning models to automatically generate brief natural language summaries of snippets. The development Transformer has led extensive use attention during model design. While existing work primarily and almost exclusively focused on static properties source related structural representations like the Abstract Syntax Tree (AST), few studies have considered human attention, that is, where programmers focus while examining comprehending code. In this...

In software engineering, interruptions during tasks can have significant implications for productivity and well-being. While previous studies investigated the effect of on productivity, to best our knowledge, no prior work has yet distinguished different types engineering activities.

Malware, or software designed with harmful intent, is an ever-evolving threat that can have drastic effects on both individuals and institutions. Neural network malware classification systems are key tools for combating these threats but vulnerable to adversarial machine learning attacks. These attacks perturb input data cause misclassification, bypassing protective systems. Existing defenses often rely enhancing the training process, thereby increasing model's robustness perturbations,...

There is a growing body of malware samples that evade automated analysis and detection tools. Malware may measure fingerprints ("artifacts") the underlying tool or environment change their behavior when artifacts are detected. While tools can mitigate to reduce exposure, such concealment expensive. However, not every sample checks for type artifact-analysis efficiency be improved by mitigating only those most likely used sample. Using insight, we propose MIMOSA, system identifies small set...

With the increasing prevalence of Web 2.0 and cloud computing, password-based logins play an increasingly important role on user-end systems. We use passwords to authenticate ourselves countless applications services. However, login credentials can be easily stolen by attackers. In this paper, we present a framework, TrustLogin, secure commodity operating TrustLogin leverages System Management Mode protect from malware even when OS is compromised. does not modify any system software in...