AI Wiki — Knowledge Base

“Marcus's repeated warnings about the 'wall of generalization' since 1998 have once again been proven true.”

“Marcus's repeated warnings about the 'wall of generalization' since 1998 have once again been proven true.” Marcus氏が1998年から繰り返す'汎化の壁'の警告。またも証明された。完璧なAIを待つより、今の限界を熟知して使いこなすチームが勝つ。少人数ゆえの意思決定の速さとリスク許容度が

safetygary-marcus--x

12 May 2026

safety

Interactive Inverse Reinforcement Learning of Interaction Scenarios via Bi-level Optimization

Internalizing Safety Understanding in Large Reasoning Models via Verification

Investigating Anisotropy in Visual Grounding under Controlled Counterfactual Perturbations

Is Class Signal Clustered or Routed in Task-Induced Implicit Neural Representation Weight Spaces?

Ister: Linear Transformer for Efficient Multivariate Time Series Forecasting

Iterative Critique-and-Routing Controller for Multi-Agent Systems with Heterogeneous LLMs

KeyframeFace: Language-Driven Facial Animation via Semantic Keyframes

Language Conditioned Multi-Finger Dexterous Manipulation Enabled by Physical Compliance and Switching of Controllers

LAQuant: A Simple Overhead-free Large Reasoning Model Quantization by Layer-wise Lookahead Loss

Large Language Models for Sequential Decision-Making: Improving In-Context Learning via Supervised Fine-Tuning

Latent Personality Alignment: Improving Harmlessness Without Mentioning Harms

LaWM: Least Action World Models for Long-Horizon Physical Consistency from Visual Observations

Learning Approximate Nash Equilibria in Cooperative Multi-Agent Reinforcement Learning via Mean-Field Subsampling

Learning the Preferences of a Learning Agent

Learning to Align Generative Appearance Priors for Fine-grained Image Retrieval

Learning to Compress Time-to-Control: A Reinforcement Learning Framework for Chronic Disease Management

Learning to Explore: Scaling Agentic Reasoning via Exploration-Aware Policy Optimization

Learning to Stay Safe: Adaptive Regularization Against Safety Degradation during Fine-Tuning

Learning When to Jump for Off-road Navigation

Learning When to Stop: Selective Imitation Learning Under Arbitrary Dynamics Shift

Let the Target Select for Itself: Data Selection via Target-Aligned Paths

LiLAW: Lightweight Learnable Adaptive Weighting to Learn Sample Difficulty & Improve Noisy Training

Liouville PDE-based sliced-Wasserstein flow

LLM Advertisement based on Neuron Auctions

LLM-Agnostic Semantic Representation Attack

LLM-Driven Performance-Space Augmentation for Meta-Learning-Based Algorithm Selection

Long-Horizon Q-Learning: Accurate Value Learning via n-Step Inequalities

LoopVLA: Learning Sufficiency in Recurrent Refinement for Vision-Language-Action Models

M^3: Reframing Training Measures for Discretized Physical Simulations

Machine Unlearning on Pre-trained Models by Residual Feature Alignment Using LoRA

MAG-VLAQ: Multi-modal Aerial-Ground Query Aggregation for Cross-View Place Recognition

Make Each Token Count: Towards Improving Long-Context Performance with KV Cache Eviction

MapFormer: Self-Supervised Learning of Cognitive Maps with Input-Dependent Positional Embeddings

“Marcus's repeated warnings about the 'wall of generalization' since 1998 have once again been proven true.”

MARLaaS: Multi-Tenant Asynchronous Reinforcement Learning as a Service

MARS-SQL: A multi-agent reinforcement learning framework for Text-to-SQL

MASS-DPO: Multi-negative Active Sample Selection for Direct Policy Optimization

Mechanism Design Is Not Enough: Prosocial Agents for Cooperative AI

MedFL-Stress: A Systematic Robustness Evaluation of Federated Brain Tumor Segmentation under Cross-Hospital MRI Appearance Shift

Mem-W: Latent Memory-Native GUI Agents

Memory Inception: Latent-Space KV Cache Manipulation for Steering LLMs

Mental Health AI Safety Claims Must Preserve Temporal Evidence

MePo: Meta Post-Refinement for Rehearsal-Free General Continual Learning

Meta offers to give rival AI chatbots free access to WhatsApp for a month while it discusses commitments with EU antitrust regulators to address their concerns (Foo Yun Chee/Reuters)

Meta-reinforcement learning with minimum attention

Metropolis-Adjusted Diffusion Models

Mid-Training with Self-Generated Data Improves Reinforcement Learning in Language Models

Mismatch-Aware Adaptive Constraint Tightening for Bicycle-Model Trajectory Optimization