cs papers | Gist.Science

Overhead-Aware Multilayer Graph Learning for Network-Managed Multi-Attack Detection in RPL-Based Low-Power IoT Networks

This paper proposes an overhead-aware multilayer graph learning framework (Attn-ML-GCN) that models RPL-based IoT networks as four-layer graphs to effectively detect multiple attack families by capturing structural and temporal attack propagation, achieving significantly higher detection accuracy than flat-feature baselines while providing interpretable insights into network overhead impacts.

Muhammad Kazim2026-08-25

💻 computer science

PVApp: A Cloud-Backed Mobile Framework for Real-Time Consumer Product Verification in Ghana

This paper presents PVApp, a cloud-backed mobile framework utilizing Firebase and QR-code scanning to enable real-time consumer product verification, approval status checking, and incident reporting in Ghana, successfully validated through functional testing to address challenges posed by counterfeit and substandard goods.

Ibrahim Tanko, Nana Kofi Ahoi Appiah-Badu, Siddique Abubakr Muntaka, Ajugu Simon2026-08-25

💻 computer science

Self-Optimizing Enterprise RAG+SQL Knowledge Agents: Unified Retrieval, Schema-Aware Query Translation, Event Graph Reasoning, and Adaptive Model Routing

This paper presents a Self-Optimizing Enterprise RAG+SQL Knowledge Agent, a middleware architecture that unifies schema-aware SQL translation, hybrid document and graph retrieval, and adaptive model routing to treat complex enterprise question answering as a self-optimizing control problem over heterogeneous knowledge tools.

Harshil Lodhiya2026-08-25

💻 computer science

Specification-first convergence with an AI coding agent: a case study of dismantling a core architectural invariant across 189 files in a 717k-line codebase with no test oracle and no human code review

This paper presents a case study demonstrating that an AI coding agent, operating under a specification-first protocol with no human code review or pre-existing test oracle, successfully dismantled a core architectural invariant across 189 files in a 717k-line TypeScript codebase by iteratively refining a formal specification and correcting 201 defects over three days for $2,430.

Joël Abenhaïm2026-08-25

💻 computer science

Risk Aware Prompt-Guided CNN–LLM Framework for Brain MRI Diagnostic Report Generation

This paper proposes a risk-aware, prompt-guided CNN–LLM framework that integrates deep visual perception with expert radiologist feedback to generate accurate, clinically interpretable, and bias-resistant brain MRI diagnostic reports, significantly outperforming conventional methods in report quality and consistency.

Senthamilselvi M., Kalavathi Palanisamy2026-08-25

💻 computer science

A ModernBERT-Based Web Application Firewall for Multi-Class HTTP Attack Detection

This paper proposes a ModernBERT-based Web Application Firewall that mitigates dataset label leakage caused by endpoint bias through an endpoint-agnostic preprocessing pipeline, achieving over 99.7% accuracy and low latency in detecting multi-class HTTP attacks without relying on request paths.

Saravanan C, Rohith K, Santhosh S, Sanjay Prabhu S, Majidha Fathima K.M2026-08-25

💻 computer science

Adversarial Contingency Auctions: Strategic Multi-Robot Task Allocation under Inconsistent Beliefs and Adversarial Path Blockages

This paper introduces Adversarial Contingency Auctions (ACA), a decentralized multi-robot task allocation framework that integrates contingency branching trees and Bayesian belief modeling to enable autonomous, localized recovery from adversarial path blockages without requiring global re-auctions, thereby significantly improving task completion rates and reducing communication overhead in dynamic, uncertain environments.

Kumar Mrinal2026-08-25

💻 computer science

Leveraging Large Language Models for Assessments Grading and Personalised Feedback Generation: A Case of Software Engineering

This study evaluates the effectiveness of LLMs (ChatGPT-4 and Gemini 2.5) in grading Software Engineering assessments and generating personalized feedback, finding that while they produce high-quality feedback and preliminary scores, their limited correlation with instructor grading suggests they are best suited for a hybrid model that supports rather than replaces human instructors.

Mamoona Tasadduq, Fakhra Amjad, Javed Ali Khan, Hafsa Shareef Dar2026-08-25

💻 computer science

Machine Learning Approaches for Hourly Emergency Department Patient Arrival Forecasting: A Multi-Horizon Comparison with Operational Benchmarking at a Norwegian Hospital

This study demonstrates that machine learning models, particularly a Direct LSTM+Optuna architecture that avoids the "frozen-feature" problem inherent in recursive autoregressive approaches, significantly outperform Norway's operational calendar-based baseline for hourly emergency department patient arrival forecasting across both short and extended planning horizons.

Md Ariful Islam Foysal, Gustav Trygve Siqueland, Vetle Ellingsen Hauge, Rasmus Rimestad, Vimala Nunavath, Bjørn-Jostein (…)2026-08-25

💻 computer science

Ecological Proportionality in Generative AI: The Ethics of Marginal Capability and Environmental Sufficiency

This article proposes an "Ecological Capability Proportionality Test" and a corresponding governance framework to ethically guide the selection of generative AI models by requiring that developers avoid significantly higher environmental costs unless the marginal ecological burden is justified by a proportionate, context-specific increase in capability.

Prudvi Saisaran Ponduru, Pavani Priya Vyshnavi Nandanavanam, Sai Kesav Kumar Ponduru2026-08-25

← Previous Next →