Association for the Advancement of Artificial Intelligence: AAAI Publications
Not a member yet
26155 research outputs found
Sort by
Online Learning-Based Android Malware Detection Using API Call Graphs and Drift Detection: A Comparative Study
No full textThe rapid growth and complexity of Android applications have made the platform a serious target for cybercriminals, posing substantial risks to mobile security and user data. Traditional malware detection models, although they have shown promise, can hardly be applied at run-time since they cannot adapt quickly enough to new malware variants and evolving attack methods. Such models, trained on preexisting data, suffer from performance degradation due to concept drift, where data distributions change over time as malware evolves. This paper presents an Online Learning-Based Android Malware Detection framework that systematically pairs various drift detection algorithms—such as ADWIN, DDM, and EDDM—with various machine learning models to identify the most effective combinations for maintaining detection accuracy in real-time. Our best-performing model achieved an accuracy of up to 96.01%
Advancing Sign Language Recognition: A YOLO v.11-Based Deep Learning Framework for Alphabet and Transactional Hand Gesture Detection
No full textSign language recognition is an essential tool that facilitates communication for those with hearing and speech disabilities. Conventional recognition techniques frequently encounter challenges in real-time performance, resilience, and accuracy owing to fluctuations in hand positions, backdrops, and lighting conditions. This paper presents a YOLOv11-based deep learning system for recognising American Sign Language (ASL), concentrating on both alphabetic and transactional hand motions to mitigate existing constraints. The model is engineered to function in real-time while ensuring high precision and resilience across varied contexts. The methodology adheres to a systematic pipeline, commencing with dataset gathering and pre-processing, which include image augmentation, normalisation, and scaling to guarantee model generalisation. The YOLOv11 architecture utilises an improved backbone, neck, and detecting head for effective feature extraction and classification. Training is enhanced by the utilisation of the AdamW optimiser, a meticulously adjusted learning rate, and a loss function that integrates box loss, classification loss, and distribution focal loss (DFL). Performance is assessed using precision, recall, mean Average Precision (mAP), and inference rate to guarantee the model's accuracy and efficiency. Experimental findings indicate that the suggested model attains 95.4% precision, 94.8% recall, and 98.1% mean Average Precision (mAP), markedly surpassing conventional methods. The amalgamation of GRAD-CAM with occlusion sensitivity significantly improves model interpretability. This research offers a robust and scalable approach for real-time sign language detection, facilitating enhanced accessibility in communication technologies, assistive devices, and interactive systems
Employing Computer Vision on a Smartphone to Help the Visually Impaired Cross the Road
No full textThis paper presents a smartphone-based hybrid computer vision system designed to assist visually impaired (VI) individuals in safely navigating pedestrian crosswalks. Existing assistive technologies often depend on controlled crossings and require external hardware, limiting their usability in diverse real-world scenarios. In contrast, this system leverages a standard smartphone camera to detect vehicles and recognize pedestrian traffic lights in real time. The proposed framework integrates two lightweight YOLOv11 models—one for vehicle detection and another for pedestrian traffic light classification—alongside MiDaS v2.1 for monocular depth estimation. These models were trained on public datasets (KITTI and blind-assist1), optimized using TensorFlow Lite, and deployed as two Android applications providing auditory feedback for real-time guidance. Performance evaluations demonstrate high accuracy in object detection and reliable depth estimation under various conditions. Usability testing further confirms the practicality of the system in live environments. By combining accessibility, mobility, and context-aware scene understanding, this work offers a low-cost, deployable alternative for improving independent mobility in the VI community
Howard's Policy Iteration is Subexponential for Deterministic Markov Decision Problems with Rewards of Fixed Bit-size and Arbitrary Discount Factor
No full textHoward's Policy Iteration (HPI) is a classic algorithm for solving Markov Decision Problems (MDPs). HPI uses a "greedy" switching rule to update from any non-optimal policy to a dominating one, iterating until an optimal policy is found. Despite its introduction over 60 years ago, the best-known upper bounds on HPI's running time remain exponential in the number of states---indeed even on the restricted class of MDPs with only deterministic transitions (DMDPs). Meanwhile, the tightest lower bound for HPI for MDPs with a constant number of actions per state is only linear. In this paper, we report a significant improvement: a subexponential upper bound for HPI on DMDPs, which is parameterised by the bit-size of the rewards, while independent of the discount factor. The same upper bound also applies to DMDPs with only two possible rewards (which may be of arbitrary size)
Quality Diversity for Variational Quantum Circuit Optimization
No full textOptimizing the architecture of variational quantum circuits (VQCs) is crucial for advancing quantum computing (QC) towards practical applications. Current methods range from static ansatz design and evolutionary methods to machine learned VQC optimization, but are either slow, sample inefficient or require infeasible circuit depth to realize advantages. Quality diversity (QD) search methods combine diversity-driven optimization with user-specified features that offer insight into the optimization quality of circuit solution candidates. However, the choice of quality measures and the representational modeling of the circuits to allow for optimization with the current state-of-the-art QD methods like covariance matrix adaptation (CMA), is currently still an open problem. In this work we introduce a directly matrix-based circuit engineering, that can be readily optimized with QD-CMA methods and evaluate heuristic circuit quality properties like expressivity and gate-diversity as quality measures. We empirically show superior circuit optimization of our QD optimization w.r.t. speed and solution score against a set of robust benchmark algorithms from the literature on a selection of NP-hard combinatorial optimization problems
Dialogic Learning in Child-Robot Interaction: A Hybrid Approach to Personalized Educational Content Generation
No full textDialogic learning fosters motivation and deeper understanding in education through purposeful and structured dialogues. Foundational models offer a transformative potential for child-robot interactions, enabling the design of personalized, engaging, and scalable interactions. However, their integration into educational contexts presents challenges in terms of ensuring age-appropriate and safe content and alignment with pedagogical goals. We introduce a hybrid approach to designing personalized educational dialogues in child-robot interactions. By combining rule-based systems with LLMs for selective offline content generation and human validation, the framework ensures educational quality and developmental appropriateness. We illustrate this approach through a project aimed at enhancing reading motivation, in which a robot facilitated book-related dialogues
EWMoE: An Effective Model for Global Weather Forecasting with Mixture-of-Experts
No full textWeather forecasting is a crucial task for meteorologic research, with direct social and economic impacts. Recently, data-driven weather forecasting models based on deep learning have shown great potential, achieving superior performance compared with traditional numerical weather prediction methods. However, these models often require massive training data and computational resources. In this paper, we propose EWMoE, an effective model for accurate global weather forecasting, which requires significantly less training data and computational resources. Our model incorporates three key components to enhance prediction accuracy: 3D absolute position embedding, a core Mixture-of-Experts (MoE) layer, and two specific loss functions. We conduct our evaluation on the ERA5 dataset using only two years of training data. Extensive experiments demonstrate that EWMoE outperforms current models such as FourCastNet and ClimaX at all forecast time, achieving competitive performance compared with the state-of-the-art models Pangu-Weather and GraphCast in evaluation metrics such as Anomaly Correlation Coefficient (ACC) and Root Mean Square Error (RMSE). Additionally, ablation studies indicate that applying the MoE architecture to weather forecasting offers significant advantages in improving accuracy and resource efficiency
MOL-Mamba: Enhancing Molecular Representation with Structural & Electronic Insights
No full textMolecular representation learning plays a crucial role in various downstream tasks, such as molecular property prediction and drug design. To accurately represent molecules, Graph Neural Networks (GNNs) and Graph Transformers (GTs) have shown potential in the realm of self-supervised pretraining. However, existing approaches often overlook the relationship between molecular structure and electronic information, as well as the internal semantic reasoning within molecules. This omission of fundamental chemical knowledge in graph semantics leads to incomplete molecular representations, missing the integration of structural and electronic data. To address these issues, we introduce MOL-Mamba, a framework that enhances molecular representation by combining structural and electronic insights. MOL-Mamba consists of an Atom & Fragment Mamba-Graph (MG) for hierarchical structural reasoning and a Mamba-Transformer (MT) fuser for integrating molecular structure and electronic correlation learning. Additionally, we propose a Structural Distribution Collaborative Training and E-semantic Fusion Training framework to further enhance molecular representation learning. Extensive experiments demonstrate that MOL-Mamba outperforms state-of-the-art baselines across eleven chemical-biological molecular datasets
An Evaluation Framework for Product Images Background Inpainting Based on Human Feedback and Product Consistency
No full textIn product advertising applications, the automated inpainting of backgrounds utilizing AI techniques in product images has emerged as a significant task. However, the techniques still suffer from issues such as inappropriate background and inconsistent product in generated product images, and existing approaches for evaluating the quality of generated product images are mostly inconsistent with human feedback causing the evaluation for this task to depend on manual annotation. To relieve the issues above, this paper proposes Human Feedback and Product Consistency (HFPC), which can automatically assess the generated product images based on two modules. Firstly, to solve inappropriate backgrounds, human feedback on 44,000 automated inpainting product images is collected to train a reward model based on multi-modal features extracted from BLIP and comparative learning. Secondly, to filter generated product images containing inconsistent products, a fine-tuned segmentation model is employed to segment the product of the original and generated product images and then compare the differences between the above two. Extensive experiments have demonstrated that HFPC can effectively evaluate the quality of generated product images and significantly reduce the expense of manual annotation. Moreover, HFPC achieves state-of-the-art (96.4% in precision) in comparison to other open-source visual-quality-assessment models
Defense Against Model Stealing Based on Account-Aware Distribution Discrepancy
No full textMalicious users attempt to replicate commercial models functionally at low cost by training a clone model with query responses. It is challenging to timely prevent such model-stealing attacks to achieve strong protection and maintain utility. In this paper, we propose a novel non-parametric detector called Account-aware Distribution Discrepancy (ADD) to recognize queries from malicious users by leveraging account-wise local dependency. We formulate each class as a Multivariate Normal distribution (MVN) in the feature space and measure the malicious score as the sum of weighted class-wise distribution discrepancy. The ADD detector is combined with random-based prediction poisoning to yield a plug-and-play defense module named D-ADD for image classification models. Results of extensive experimental studies show that D-ADD achieves strong defense against different types of attacks with little interference in serving benign users for both soft and hard-label settings