cs papers | Gist.Science

Image Tampering Region Localization Method Based on Pseudo Label Driver and Dynamic Multi branch Decoding

This paper proposes a pseudo-label-driven image tampering localization method featuring a tamper-type-aware dynamic multi-branch decoder that effectively addresses heterogeneous forensic traces by integrating multi-scale features, source–target correspondence supervision for copy-move forgery, and adaptive cue fusion to achieve high accuracy and generalization across diverse datasets.

Yanting Li, Xiaojian Xue, Weiyi Lin, Shouxin He2026-09-15

💻 computer science

Resource-Aware Small-UAV Perception under Annotation and Compute Constraints

This paper proposes and evaluates a resource-aware perception strategy for small UAV detection that combines confidence-guided model routing and difficulty-ranked replay to effectively optimize performance under simultaneous constraints on inference-time computation and annotation budgets.

Xufeng Zhou, Jun Li, Hua Hu, Yan Liu, Xiaodong Wang, Kebin Chen, Chao Jiang, Quantong Lv, Xuping Jiang2026-09-15

💻 computer science

Cross-cohort evaluation of thyroid ultrasound deep learning across different outcome definitions: a duplicate-controlled benchmark

This study demonstrates that frozen thyroid ultrasound deep learning models exhibit significantly varying performance across different cohorts and outcome definitions, highlighting the critical need for explicit reporting of label provenance and cautious interpretation of cross-cohort benchmarks.

Behnam Kiani Kalejahi, Mohammad Javad Rajabi2026-09-15

💻 computer science

Lightweight Prompt-Based Enhancement for Missing-Modality Multimodal Sentiment Analysis

This paper proposes a lightweight, prompt-based framework comprising three synergistic modules—Authentic Feature Enhancement, Reliability-Aware Fusion, and Content-Guided Distribution Adapter—to effectively address modality missing in multimodal sentiment analysis by recovering fine-grained features, dynamically adjusting fusion weights, and correcting distribution shifts.

Juan Liu, Jinping Liu, Hang Zhong, Shikang He2026-09-15

💻 computer science

FL-GWO: A Federated Learning and Grey Wolf Optimization Framework for Blockchain-Assisted Trust Management in Decentralized IoT-Edge Systems

This paper proposes FL-GWO, a novel framework that integrates federated learning with Grey Wolf Optimization and blockchain technology to enable privacy-preserving, adaptive, and robust trust management in decentralized IoT-edge systems by dynamically optimizing trust parameters through swarm intelligence.

Deepshikha Arya, Virendra Kumar Tiwari, Neelu Singh, Seema Joshi2026-09-15

💻 computer science

The invisible majority: linking a national researcher registry to OpenAlex reveals Peru's uncertified artificial intelligence workforce

By linking Peru's national researcher registry (RENACYT) to OpenAlex, this study reveals that the country's certified artificial intelligence workforce captures only a quarter of its actual AI authors, exposing a significantly larger uncertified majority and highlighting regional and gender disparities in certification despite high productivity outside the capital.

Milton Vladimir Mamani Calisaya, Charles Ignacio Mendoza Mollocondo, Erardo Espinoza Coaquira, Jesus Pari Flores, Sandra (…)2026-09-15

💻 computer science

Microservice Root Cause Localization Based on Bi-Variate Graph Variational Autoencoder with Counterfactual-Inspired Recovery Scoring

This paper proposes BVC-RCA, a microservice root cause localization model that integrates a Parameter-Enhanced Heterogeneous Trace-Log Graph, a Bi-variate Graph Variational Autoencoder, and a counterfactual-inspired recovery scoring mechanism to effectively distinguish true root causes from cascading victims by unifying multi-source observability data and measuring node-level recovery contributions.

Jian Feng, Jiang Zheng, Haizheng Duan, Jiawei Liu2026-09-15

💻 computer science

Human Fall Detection Using Deep Learning Methods: A Multimodal Audio-Video Approach

This study proposes a multimodal deep learning framework that fuses audio and video features using decision-level stacking to achieve a 98% accuracy in human fall detection, significantly outperforming single-modality models and other fusion strategies.

Hamid Ghous, Qandeel Asghar, Numan Asghar, Syed Adil Hussain Shah, Syed Taimoor Hussain Shah, Marco Agostino Deriu2026-09-15

💻 computer science

Adversarially Robust and Semantically Aware Phishing Detection using Calibrated Multimodal Learning

This paper presents a robust, semantically aware multimodal phishing detection framework that fuses URL, text, and visual features with calibrated confidence and adversarial training to significantly outperform unimodal baselines while maintaining high accuracy and reliability under deceptive attacks.

Dinesh Patil, Madhuri Gedam, Chhaya dhavale2026-09-15

💻 computer science

Runtime Risk Detection and Control for AI Agents and LLM Applications

This article presents an artefact-based case study of AgentGuard AI v2.4.0, a research prototype that operationalizes runtime risk detection and control mechanisms for AI agents, while highlighting its utility as an inspectable governance instrument and identifying specific reproducibility defects that preclude claims of production effectiveness.

Deepak Kumar Gour, Sushma Agrawal2026-09-15

← Previous Next →