Skip to content

usdaed

  • Science
  • Who is Denis Avetissian?

Science

Spotting Power Grid Issues with AI: A New Approach to Anomaly Detection

15.01.2026 by qfx

A system designed for anomaly detection leverages adversarial training-specifically, a generative adversarial network coupled with a long short-term memory network-trained exclusively on normal operational windows, and subsequently employs frozen network weights during a test phase to optimize within the latent space, ultimately yielding anomaly scores for classification-a process acknowledging that all systems inevitably deviate from nominal behavior and that discerning those deviations is crucial, not for prevention, but for understanding the nature of decay.

Researchers are leveraging the power of generative adversarial networks and long short-term memory networks to improve the accuracy and reliability of identifying unusual energy consumption patterns.

Categories Science

Seeing Through Time: Can AI Decode Ancient Roman Coin Imagery?

15.01.2026 by qfx

The convolutional neural network, when identifying key features, demonstrated a pronounced focus on the upper-left quadrant of images, frequently misinterpreting depictions of crosses carried by angels - often found on coins bearing shield imagery on their reverse - as shields themselves.

A new study investigates whether modern computer vision techniques, specifically Vision Transformers and Convolutional Neural Networks, can accurately identify motifs on ancient Roman coins.

Categories Science

The Silent Breach: AI Versus AI in Industrial Control Systems

15.01.2026 by qfx

The study demonstrates the vulnerability of field networks to sophisticated attacks leveraging compromised Programmable Logic Controllers (PLCs), Remote I/O devices, or low-level (L0L\_{0}) communication manipulation, highlighting a critical need for robust cybersecurity measures in industrial control systems.

New research reveals a sophisticated adversarial strategy leveraging artificial intelligence to compromise critical infrastructure undetected by existing AI-powered security measures.

Categories Science

Intelligent Skies: AI Powers Resource Allocation for Satellite Networks

15.01.2026 by qfx

A proposed framework leverages large language models to translate environmental states and operator goals into strategic action labels, embedding these strategies as conditioning vectors within a single-head attention layer to sculpt the resulting reward signal-a process wherein the system doesn’t simply react to conditions, but actively shapes its own incentive landscape.

A new approach combines the power of large language models with deep reinforcement learning to optimize resource management in Low Earth Orbit satellite communications.

Categories Science

Beyond Basic Prompts: Elevating AI’s Emotional Intelligence

14.01.2026 by qfx

The comparison reveals a baseline against which one-shot irony detection-using the gemini-flash1.5 model-is measured, highlighting the system's capacity to discern nuanced linguistic shifts even within a single instance of expression.

New research reveals how carefully crafted prompts can unlock significant gains in large language models’ ability to understand sentiment and detect nuance like irony.

Categories Science

Why Deep Learning Can Still Surprise Us

14.01.2026 by qfx

The model’s learning progression-from initial fitting of clean data, through a phase of noisy data integration and signal separation, to a state of benign overfitting-demonstrates that large activations compress input patterns, enabling sparse networks to effectively learn essential data characteristics and ultimately achieve double descent-a second decrease in test loss-despite the presence of noise.

New research reveals a hidden mechanism behind the surprisingly robust performance of deep neural networks, even when overfitted to noisy data.

Categories Science

Why AI Writing Detectors Often Fail

14.01.2026 by qfx

New research reveals the surprisingly fragile foundations of algorithms designed to identify text written by artificial intelligence.

Categories Science

When RL Gets Stuck: Solving Challenges in Dynamic Environments

14.01.2026 by qfx

An agent navigating a multiplicative dynamic optimizes wealth by selecting between safe and risky actions, where repetitions in action selection-returning updated wealth to the state-manage non-ergodic contexts, and differing optimization strategies-based on expected values versus growth rates-predict distinct indifference points between the two actions, [latex]p_E[/latex] and [latex]p_T[/latex], respectively.

New research reveals why standard reinforcement learning agents falter in constantly changing situations and proposes a surprisingly simple fix.

Categories Science

Beyond Accuracy: Stress-Testing Deepfake Detection

14.01.2026 by qfx

The study contrasts conventional deepfake detector evaluation-focused solely on performance and robustness-with a novel framework that integrates and quantifies four foundational pillars of reliability, offering a more comprehensive assessment.

A new study reveals that current deepfake detection methods often falter when faced with real-world conditions and subtle manipulations.

Categories Science

Outsmarting the Bots: AI-Powered Defense for Memecoin Traders

14.01.2026 by qfx

A multi-agent system, driven by large language models, orchestrates a workflow designed for automated copy trading of meme coins.

A new multi-agent system leverages the power of large language models to detect and mitigate manipulation in the volatile world of memecoin copy trading.

Categories Science
Older posts
Newer posts
← Previous Page1 … Page16 Page17 Page18 … Page87 Next →
© 2026 usdaed • Built with GeneratePress