4159 papers
This empirical study demonstrates that while fusing Strace and DroidBox runtime features significantly improves predictive performance in dynamic Android malware detection, behavioral richness and predictive utility remain distinct properties, as execution duration and sequence ordering effects do not consistently translate into better discrimination.
This paper introduces HyperCookies, a publicly available hyperspectral imaging dataset of 160 cookies with varied cocoa contents and topping arrangements, designed to serve as a challenging benchmark for developing machine learning methods that effectively combine spectral and spatial information for classification.
This study improves English–Urdu neural machine translation by compiling a large-scale parallel corpus and demonstrating that optimizing subword tokenization strategies (specifically Subword-NMT BPE and SentencePiece-BPE) combined with Dynamic Position Encoding effectively addresses vocabulary sparsity and word-order divergence to outperform baseline models.
This paper introduces SCOPE, a stateful FaaS programming model and execution platform for the edge-to-cloud continuum that utilizes dynamic session management to balance low-latency proximity with resource availability, thereby overcoming the stateless limitations of current serverless edge solutions.
The paper introduces GANGSTER, a generative adversarial network framework that enhances off-policy reinforcement learning algorithms like TD3 and SAC by training structured subgoals to significantly improve sample efficiency, stability, and final performance in complex continuous-control environments.
This paper precisely classifies the complexity of Boolean formula isomorphism within the polynomial hierarchy by establishing a bidirectional many-one equivalence between canonical formula isomorphism and graph isomorphism, while demonstrating that general formula isomorphism resides at the second level () due to the combination of coNP-complete equivalence testing and the residual graph isomorphism core.
This study finds that adding structured metadata to text-based classifiers offers no consistent incremental benefit for university mental health risk triage, as message text alone outperforms structured-only models and combined approaches, though the role-play nature of the data and lack of autonomous-level performance necessitate further validation on authentic counselling records.
This paper proposes SECURE, a domain-independent six-stage framework that integrates a Random Forest-based Intrusion Detection System with SHAP-driven Explainable AI to effectively detect and interpret integrity attacks, such as spoofing and layering, within high-frequency trading data streams.
This paper proposes a Differential Evolution-initialized Hybrid Cuckoo Search Algorithm for feature selection in CoWIN Twitter sentiment analysis, demonstrating that the method effectively reduces data dimensionality by 48.40% while achieving high classification accuracy (82.90%) and F1 scores (73.51%) compared to other optimization techniques.
This paper presents a PyTorch-based multitask GRU model that successfully integrates channel disturbance detection, cause classification, and QBER forecasting for simulated decoy-state BB84 QKD links, achieving high accuracy in controlled environments while acknowledging limitations in generalizing to physical systems and unseen disturbance types.