Science – Page 17

Taming Market Jumps with Reinforcement Learning

11.12.2025 by qfx

$The convergence of parameters $\mu$, $\sigma$, and $\delta$ indicates a refined optimization process, suggesting the model effectively narrows its uncertainty and stabilizes towards a defined solution through iterative adjustments to these key variables.$

A new approach combines reinforcement learning and equilibrium concepts to optimize investment portfolios even when faced with sudden, unpredictable market shifts.

Mining Materials Data with AI: Progress and Persistent Challenges

11.12.2025 by qfx

A new analysis assesses how well artificial intelligence tools can automatically pull crucial data from the ever-growing body of materials science research.

When AI Bargains: The Flaws in Rational Negotiation

11.12.2025 by qfx

$Despite advancements in reasoning capabilities, frontier models-specifically Claude 4.5 Sonnet ($\rho\approx 0.78$) and Gemini 2.5 Pro ($\rho\approx 0.91$)-exhibit a strong susceptibility to numerical anchoring bias, as demonstrated by the tight clustering of final prices around initial proposals in self-play simulations, suggesting that even sophisticated systems are bound by cognitive heuristics rooted in their initial conditions.$

Despite impressive gains in artificial intelligence, new research reveals that even the most advanced language models struggle with fundamental biases and strategic inconsistencies during complex negotiations.

Securing the Blockchain: A New Approach to Smart Contract Vulnerability Detection

11.12.2025 by qfx

BugSweeper establishes a robust vulnerability detection system by transforming contract code into Function-Level Abstract Syntax Graphs (FLAGs)-augmented with control-flow and data-flow information-and subsequently analyzing these graphs with a two-stage Graph Neural Network to identify potential security flaws.

Researchers have developed a novel framework, BugSweeper, that leverages graph neural networks to pinpoint vulnerabilities within smart contract code with greater precision.

Cosmic Dawn’s Signal: Machine Learning Cuts Through the Noise

11.12.2025 by qfx

The study constructs training datasets for analyzing the faint $21$-cm signal from the universe’s early epochs, deliberately embedding realistic foreground interference and thermal noise, with a select portion highlighted to demonstrate the range of simulated conditions against which signal detection algorithms will be tested.

Researchers are harnessing the power of machine learning to isolate the faint signals from the universe’s first stars and galaxies.

Seeing Through the Fake: New Attack Exposes AI Image Detector Weaknesses

11.12.2025 by qfx

The proposed FBA2D framework offers a novel approach to the problem, acknowledging that even the most innovative solutions inevitably contribute to future technical debt within production systems.

Researchers have demonstrated a novel method for deceiving AI-powered image detectors, highlighting critical vulnerabilities in current authentication systems.

Transformers Under the Microscope: What Graph Neural Networks Reveal

11.12.2025 by qfx

Effective resistance, a measure of how easily current dissipates across a network, varies predictably with the underlying graph structure, demonstrating that connectivity isn’t simply a topological property but a dynamic influence on energy flow-a system’s inherent susceptibility to decay is encoded within its architecture.

A new analysis frames the strengths and weaknesses of transformer models through the principles of graph neural networks, shedding light on their internal workings.

Taming Market Jumps with Reinforcement Learning

Mining Materials Data with AI: Progress and Persistent Challenges

When AI Bargains: The Flaws in Rational Negotiation

Securing the Blockchain: A New Approach to Smart Contract Vulnerability Detection

Cosmic Dawn’s Signal: Machine Learning Cuts Through the Noise

Seeing Through the Fake: New Attack Exposes AI Image Detector Weaknesses

Transformers Under the Microscope: What Graph Neural Networks Reveal

Smarter Finance: AI Agents That Explain Their Reasoning

Why Hackers Hate Uncertainty: A New Approach to Cyber Defense

Detecting Market Shifts: A Statistical Edge for Traders