cs papers | Gist.Science

Factorial Benchmarking of Exact and Finite-Shot Gradient Resolvability in a Four-Qubit Hybrid Quantum Neural Network

This study employs a complete 2³ factorial design to benchmark a four-qubit hybrid quantum neural network, revealing that while specific structural and objective interactions significantly influence exact-gradient magnitude, their effects under finite-shot conditions remain statistically ambiguous and do not yet demonstrate optimization success or hardware advantages.

Brandon Shen2026-08-11

💻 computer science

The Intent Gap: A Taxonomy of Real-User Failure Modes in Frontier AI Agents

This paper identifies a critical "intent gap" where frontier AI models fail to satisfy real user needs despite literal prompt compliance, revealing that current researcher-designed benchmarks miss these deployment-relevant failures because real users rarely verbalize their dissatisfaction, a phenomenon the author quantifies through a new taxonomy derived from mining large-scale conversation corpora.

Kanupriya Yakhmi2026-08-11

💻 computer science

HABIT: A Self-Evolving AI-Agent Harness for GUI-Based CAE Software---Application to Aerospace

This paper introduces HABIT, a self-evolving, host-agnostic AI agent harness designed to automate complex GUI-based CAE workflows in aerospace by synchronizing reliable tool calling with real-time software states and continuously adapting to user habits through a memory-driven learning system.

Shaoqi WU, Jingyi LI, Jiayi YAO, Xingyuan JU2026-08-11

💻 computer science

A Hybrid Graph Neural Network Approach for Fraud Detection in Financial Transactions

This paper proposes a novel hybrid graph neural network framework that integrates Graph Convolutional Networks, Graph Attention Networks, Graph Transformers, and wavelet-based feature enhancement to capture multi-scale structural and spectral patterns, significantly outperforming existing methods in detecting complex financial fraud with high precision and scalability.

Mohamed Hechmi2026-08-11

💻 computer science

A PINN Approach for Solving Singularly Perturbed Parabolic Partial Differential Equations with Two Small Parameters

This paper proposes two Physics-Informed Neural Network models, DeepXDE and a raw PyTorch implementation, to effectively solve singularly perturbed parabolic partial differential equations with two small parameters, addressing the limitations of standard numerical methods in capturing boundary and interior layers.

Archishman Nag, Akmal Raza, Arshad Khan2026-08-11

💻 computer science

Feature-Reduced IoT Intrusion Detection Using Clustering-Based Mutual Information and Meta-Heuristic CNN Optimization

This paper proposes a computationally efficient IoT intrusion detection pipeline that integrates a hybrid SMOTE-ENN-LOF balancing strategy, a clustering-based mutual information feature selection algorithm (BBFS), and Hunger Games Search-optimized 1D CNNs to achieve reliable attack detection with reduced feature sets and compact model architectures across diverse IoT datasets.

S KUMAR REDDY MALLIDI, Rajeswara Rao Ramisetty2026-08-11

💻 computer science

Explainable Hybrid ConvNeXt–Vision Transformer Model for Pneumonia Detection from Chest X-ray Images

This paper proposes an explainable hybrid deep learning framework combining ConvNeXt-Base with CBAM and Vision Transformer (ViT-B/16) that achieves superior pneumonia detection accuracy (94.03%) and interpretability on chest X-ray images compared to existing baseline models.

Israt Fatema Shorna, Sadek Al Prince, Aziz Misher Mishu, Gazi Mehraj Sakib, Fairuz Aman2026-08-11

💻 computer science

Entity-Level Post-Hoc Reliability Calibration for Named Entity Recognition via Multi-Source Features

This paper proposes a post-hoc entity-level reliability calibration method for Named Entity Recognition that leverages multi-source features to train a lightweight calibrator, significantly improving confidence accuracy and deployment robustness without modifying the underlying model.

Junhui Yang, Feifan Xiao2026-08-11

💻 computer science

Parkinson's Disease Drug Dose Optimization Using Multi-Objective Reinforcement Learning

This paper proposes a multi-objective reinforcement learning framework for optimizing Parkinson's disease medication dosing that generates a Pareto set of policies to explicitly model and navigate the trade-offs between motor symptom control and dyskinesia risks, outperforming clinician and random baselines while supporting transparent, patient-centered decision-making.

Ahmad Rezaie Mianroodi, Seyed-Mohammad Fereshtehnejad, Frank Rudzicz2026-08-11

💻 computer science

Engineering the First Digital Keyboard for the Zou Language: Layout Optimization and Standardized Repository Integration

This paper details the design, optimization, and standardization of the first digital keyboard for the Zou language, successfully integrating a custom character-mapping solution into the global Keyman repository to overcome hardware limitations and enable the computational preservation and digital localization of this under-resourced script.

T. Marvin Zou2026-08-11

← Previous Next →