Malware evolves rapidly, forcing machine learning (ML)-based detectors to adapt continuously. With antivirus vendors processing hundreds of thousands of new samples daily, datasets can grow to billions of examples, making full retraining impractical. Continual learning (CL) has emerged as a scalable alternative, enabling incremental updates without full data access while mitigating catastrophic forgetting. In this work, we analyze a critical yet overlooked issue in this context: security regression. Unlike forgetting, which manifests as a general performance drop on previously seen data, security regression captures harmful prediction changes at the sample level, such as a malware sample that was once correctly detected but evades detection after a model update. Although often overlooked, regressions pose serious risks in security-critical applications, as the silent reintroduction of previously detected threats in the system may undermine users' trust in the whole updating process. To address this issue, we formalize and quantify security regression in CL-based malware detectors and propose a regression-aware penalty to mitigate it. Specifically, we adapt Positive Congruent Training (PCT) to the CL setting, preserving prior predictive behavior in a model-agnostic manner. Experiments on the ELSA, Tesseract, and AZ-Class datasets show that our method effectively reduces regression across different CL scenarios while maintaining strong detection performance over time.

AI systems are gaining widespread adoption across various sectors and domains. Creating high-quality AI system requirements is crucial for aligning the AI system with business goals and consumer values and for social responsibility. However, with the uncertain nature of AI systems and the heavy reliance on sensitive data, more research is needed to address the elicitation and analysis of AI systems requirements. With the proprietary nature of many AI systems, there is a lack of open-source requirements artifacts and technical requirements documents for AI systems, limiting broader research and investigation. With Large Language Models (LLMs) emerging as a promising alternative to human-generated text, this paper investigates the potential use of LLMs to generate user stories for AI systems based on abstracts from scholarly papers. We conducted an empirical evaluation using three LLMs and generated $1260$ user stories from $42$ abstracts from $26$ domains. We assess their quality using the Quality User Story (QUS) framework. Moreover, we identify relevant non-functional requirements (NFRs) and ethical principles. Our analysis demonstrates that the investigated LLMs can generate user stories inspired by the needs of various stakeholders, offering a promising approach for generating user stories for research purposes and for aiding in the early requirements elicitation phase of AI systems. We have compiled and curated a collection of stories generated by various LLMs into a dataset (UStAI), which is now publicly available for use.

App reviews are a critical source of user feedback, offering valuable insights into an app's performance, features, usability, and overall user experience. Effectively analyzing these reviews is essential for guiding app development, prioritizing feature updates, and enhancing user satisfaction. Classifying reviews into functional and non-functional requirements play a pivotal role in distinguishing feedback related to specific app features (functional requirements) from feedback concerning broader quality attributes, such as performance, usability, and reliability (non-functional requirements). Both categories are integral to informed development decisions. Traditional approaches to classifying app reviews are hindered by the need for large, domain-specific datasets, which are often costly and time-consuming to curate. This study explores the potential of zero-shot learning with ChatGPT for classifying app reviews into four categories: functional requirement, non-functional requirement, both, or neither. We evaluate ChatGPT's performance on a benchmark dataset of 1,880 manually annotated reviews from ten diverse apps spanning multiple domains. Our findings demonstrate that ChatGPT achieves a robust F1 score of 0.842 in review classification, despite certain challenges and limitations. Additionally, we examine how factors such as review readability and length impact classification accuracy and conduct a manual analysis to identify review categories more prone to misclassification.

Neglecting non-functional requirements (NFRs) early in software development can lead to critical challenges. Despite their importance, NFRs are often overlooked or difficult to identify, impacting software quality. To support requirements engineers in eliciting NFRs, we developed a framework that leverages Large Language Models (LLMs) to derive quality-driven NFRs from functional requirements (FRs). Using a custom prompting technique within a Deno-based pipeline, the system identifies relevant quality attributes for each functional requirement and generates corresponding NFRs, aiding systematic integration. A crucial aspect is evaluating the quality and suitability of these generated requirements. Can LLMs produce high-quality NFR suggestions? Using 34 functional requirements - selected as a representative subset of 3,964 FRs-the LLMs inferred applicable attributes based on the ISO/IEC 25010:2023 standard, generating 1,593 NFRs. A horizontal evaluation covered three dimensions: NFR validity, applicability of quality attributes, and classification precision. Ten industry software quality evaluators, averaging 13 years of experience, assessed a subset for relevance and quality. The evaluation showed strong alignment between LLM-generated NFRs and expert assessments, with median validity and applicability scores of 5.0 (means: 4.63 and 4.59, respectively) on a 1-5 scale. In the classification task, 80.4% of LLM-assigned attributes matched expert choices, with 8.3% near misses and 11.3% mismatches. A comparative analysis of eight LLMs highlighted variations in performance, with gemini-1.5-pro exhibiting the highest attribute accuracy, while llama-3.3-70B achieved higher validity and applicability scores. These findings provide insights into the feasibility of using LLMs for automated NFR generation and lay the foundation for further exploration of AI-assisted requirements engineering.

The performance of active learning algorithms can be improved in two ways. The often used and intuitive way is by reducing the overall error rate within the test set. The second way is to ensure that correct predictions are not forgotten when the training set is increased in between rounds. The former is measured by the accuracy of the model and the latter is captured in negative flips between rounds. Negative flips are samples that are correctly predicted when trained with the previous/smaller dataset and incorrectly predicted after additional samples are labeled. In this paper, we discuss improving the performance of active learning algorithms both in terms of prediction accuracy and negative flips. The first observation we make in this paper is that negative flips and overall error rates are decoupled and reducing one does not necessarily imply that the other is reduced. Our observation is important as current active learning algorithms do not consider negative flips directly and implicitly assume the opposite. The second observation is that performing targeted active learning on subsets of the unlabeled pool has a significant impact on the behavior of the active learning algorithm and influences both negative flips and prediction accuracy. We then develop ROSE - a plug-in algorithm that utilizes a small labeled validation set to restrict arbitrary active learning acquisition functions to negative flips within the unlabeled pool. We show that integrating a validation set results in a significant performance boost in terms of accuracy, negative flip rate reduction, or both.

The diagnostic imaging departments are under great pressure due to a growing workload. The number of required scans is growing and there is a shortage of qualified labor. AI solutions for medical imaging applications have shown great potential. However, very few diagnostic imaging models have been approved for hospital use and even fewer are being implemented at the hospitals. The most common reason why software projects fail is poor requirement engineering, especially non-functional requirements (NFRs) can be detrimental to a project. Research shows that machine learning professionals struggle to work with NFRs and that there is a need to adapt NFR frameworks to machine learning, AI-based, software. This study uses qualitative methods to interact with key stakeholders to identify which types of NFRs are important for medical imaging applications. The study was done on a single Danish hospital and found that NFRs of type Efficiency, Accuracy, Interoperability, Reliability, Usability, Adaptability, and Fairness were important to the stakeholders. Especially Efficiency since the diagnostic imaging department is trying to spend as little time as possible on each scan.

In modern society, service robots are increasingly recognized for their wide range of practical applications. In large and crowded social spaces, such as museums and hospitals, these robots are required to safely move in the environment while exhibiting user-friendly behavior. Ensuring the safe and socially acceptable operation of robots in such settings presents several challenges. To enhance the social acceptance in the design process of service robots, we present a systematic analysis of requirements, categorized into functional and non-functional. These requirements are further classified into different categories, with a single requirement potentially belonging to multiple categories. Finally, considering the specific case of a receptionist robotic agent, we discuss the requirements it should possess to ensure social acceptance.

Negative flips are errors introduced in a classification system when a legacy model is updated. Existing methods to reduce the negative flip rate (NFR) either do so at the expense of overall accuracy by forcing a new model to imitate the old models, or use ensembles, which multiply inference cost prohibitively. We analyze the role of ensembles in reducing NFR and observe that they remove negative flips that are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits. Based on the observation, we present a method, called Ensemble Logit Difference Inhibition (ELODI), to train a classification system that achieves paragon performance in both error rate and NFR, at the inference cost of a single model. The method distills a homogeneous ensemble to a single student model which is used to update the classification system. ELODI also introduces a generalized distillation objective, Logit Difference Inhibition (LDI), which only penalizes the logit difference of a subset of classes with the highest logit values. On multiple image classification benchmarks, model updates with ELODI demonstrate superior accuracy retention and NFR reduction.

Malicious domain detection (MDD) is an open security challenge that aims to detect if an Internet domain is associated with cyber-attacks. Among many approaches to this problem, graph neural networks (GNNs) are deemed highly effective. GNN-based MDD uses DNS logs to represent Internet domains as nodes in a maliciousness graph (DMG) and trains a GNN to infer their maliciousness by leveraging identified malicious domains. Since this method relies on accessible DNS logs to construct DMGs, it exposes a vulnerability for adversaries to manipulate their domain nodes' features and connections within DMGs. Existing research mainly concentrates on threat models that manipulate individual attacker nodes. However, adversaries commonly generate multiple domains to achieve their goals economically and avoid detection. Their objective is to evade discovery across as many domains as feasible. In this work, we call the attack that manipulates several nodes in the DMG concurrently a multi-instance evasion attack. We present theoretical and empirical evidence that the existing single-instance evasion techniques for are inadequate to launch multi-instance evasion attacks against GNN-based MDDs. Therefore, we introduce MintA, an inference-time multi-instance adversarial attack on GNN-based MDDs. MintA enhances node and neighborhood evasiveness through optimized perturbations and operates successfully with only black-box access to the target model, eliminating the need for knowledge about the model's specifics or non-adversary nodes. We formulate an optimization challenge for MintA, achieving an approximate solution. Evaluating MintA on a leading GNN-based MDD technique with real-world data showcases an attack success rate exceeding 80%. These findings act as a warning for security experts, underscoring GNN-based MDDs' susceptibility to practical attacks that can undermine their effectiveness and benefits.

We propose an end-to-end deep-learning approach for automatic rigging and retargeting of 3D models of human faces in the wild. Our approach, called Neural Face Rigging (NFR), holds three key properties: (i) NFR's expression space maintains human-interpretable editing parameters for artistic controls; (ii) NFR is readily applicable to arbitrary facial meshes with different connectivity and expressions; (iii) NFR can encode and produce fine-grained details of complex expressions performed by arbitrary subjects. To the best of our knowledge, NFR is the first approach to provide realistic and controllable deformations of in-the-wild facial meshes, without the manual creation of blendshapes or correspondence. We design a deformation autoencoder and train it through a multi-dataset training scheme, which benefits from the unique advantages of two data sources: a linear 3DMM with interpretable control parameters as in FACS, and 4D captures of real faces with fine-grained details. Through various experiments, we show NFR's ability to automatically produce realistic and accurate facial deformations across a wide range of existing datasets as well as noisy facial scans in-the-wild, while providing artist-controlled, editable parameters.