Beyond Pixels: Fuzzy Logic Sharpens Brain Scan Analysis
A new approach combines the power of deep learning with intuitionistic fuzzy logic to improve the precision of brain MRI image segmentation.
Science
Beyond Neural Networks: Building AI with Geometric Principles
A new approach to artificial intelligence leverages the power of geometric algebra and Bayesian methods for more robust, verifiable, and continuously learning systems.
Smart Agents Find Equilibrium: A New Path to Stable AI Interactions
Researchers have shown that AI agents, leveraging reasonable reasoning, can independently converge on stable strategies in repeated interactions without explicit game-theoretic training.
Taming Generative Models: A Stability-Focused Approach
![The evolution of component weights [latex]\theta(t)[/latex] during the generation task demonstrates a dynamic process where individual parameters adjust over time, shaping the system’s overall behavior.](https://arxiv.org/html/2603.18022v1/control_theory.png)
Researchers are leveraging control theory and Laplace transforms to understand and mitigate the tendency of generative AI to produce unrealistic or ‘hallucinatory’ outputs.
Can Simpler Models Beat Deep Learning for Anomaly Detection?
New research challenges the assumption that complex deep learning architectures are always superior for time series anomaly detection.
Small Data, Big Insights: AI Accurately Spots Prostate Cancer
A new study demonstrates that even limited datasets of medical images can power surprisingly accurate AI detection of prostate cancer.
Smarter Clustering: Algorithms That Learn on the Fly
A new approach frames the classic k-median problem as an online learning challenge, enabling algorithms to adapt and compete with optimal solutions even as data changes.
AI’s Hidden Biases: What Your Assistant Prefers
![ChoiceEval establishes a systematic framework for generating evaluation questions and rigorously assessing entity-perception bias within artificial intelligence assistants, enabling a quantifiable understanding of potentially skewed perspectives inherent in these systems and moving beyond merely functional correctness to address foundational fairness in AI perception-a crucial step towards genuinely unbiased artificial intelligence, formalized as minimizing the divergence between expected and observed responses given a defined entity set [latex] E [/latex] and question space [latex] Q [/latex].](https://arxiv.org/html/2603.18300v1/x1.png)
New research reveals that AI assistants consistently favor certain brands and cultures, raising questions about fairness and representation in automated recommendations.
Beyond the Black Box: Illuminating Prompt Optimization
New research tackles the interpretability challenges of automatically refining prompts for large language models, revealing why some methods fail and offering a path to more reliable performance.
The Feedback Loop of Bias: How Predictive Policing Amplifies Racial Disparities
New research reveals that AI-powered predictive policing systems, even with attempts at data correction, can worsen existing biases and lead to significantly unequal outcomes.