Neural Networks

LLM Research

New Method Internalizes LLM Reasoning Through Latent Actions

Researchers propose a novel approach to improve LLM reasoning by discovering and replaying latent actions, potentially reducing inference costs while maintaining reasoning quality.

LLM Research

Activation Steering: How Reasoning-Critical Neurons Improve LLM R

New research identifies specific neurons responsible for reasoning in LLMs and demonstrates how transferring their activation patterns can significantly improve inference reliability across models.

Neural Networks

Differential Geometry Unlocks Neural Network Information Flow

New research applies differential geometry to analyze how information propagates through neural networks, offering mathematical tools to understand deep learning architectures at a fundamental level.

AI Research

AI Memory Systems: How Close Are We to Human Hippocampus?

New research examines the gap between AI memory architectures and the human hippocampus, exploring how neuroscience insights could transform machine learning systems.

Deep Learning

NOVAK: A Unified Adaptive Optimizer for Deep Neural Networks

New research introduces NOVAK, a unified framework that bridges popular adaptive optimizers like Adam and AdaGrad, potentially improving training efficiency for deep learning models.

Explainable AI

xDNN(ASP): Logic-Based Explainability for Neural Networks

New research combines deep neural networks with Answer Set Programming to generate human-readable explanations for AI decisions, advancing interpretability crucial for detection systems.

Machine Learning

Gradient Descent Explained: The Core Algorithm Behind AI

Understanding gradient descent is essential to grasping how neural networks learn. This foundational optimization algorithm powers everything from deepfake generators to detection systems.

AI Safety

Formal Verification Breakthrough for Early-Exit Neural Networks

New research bridges efficiency and safety by developing formal verification methods for neural networks with early exits, enabling mathematically proven safety guarantees for adaptive AI systems.

LLM Interpretability

Brain-Grounded Axes: Reading and Steering LLM Internal States

New research maps LLM internal representations to brain-derived axes, enabling interpretable reading and targeted steering of model behavior without fine-tuning.

LLM Agents

ABBEL: Language-Based Belief Bottlenecks Improve LLM Agents

New research introduces ABBEL, an architecture that constrains LLM agents to act through explicit belief states expressed in natural language, improving interpretability and decision-making in complex environments.

AI Security

ArcGen: Cross-Architecture Backdoor Detection for Neural Networks

New research introduces ArcGen, a framework that generalizes neural backdoor detection across diverse model architectures without retraining, addressing critical AI security vulnerabilities.

Neural Networks

Training Neural Networks Without Backpropagation

New research proposes training graph-based neural networks using few-shot learning without traditional backpropagation, potentially revolutionizing how AI models are trained.