AI Infrastructure

GPUs, training clusters, MLOps, and deployment

Top This Week

[2604.01989] Attention at Rest Stays at Rest: Breaking Visual Inertia for Cognitive Hallucination Mitigation
Llms

[2604.01989] Attention at Rest Stays at Rest: Breaking Visual Inertia for Cognitive Hallucination Mitigation

Abstract page for arXiv paper 2604.01989: Attention at Rest Stays at Rest: Breaking Visual Inertia for Cognitive Hallucination Mitigation

arXiv - AI · 4 min ·
[2512.18809] FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation
Machine Learning

[2512.18809] FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation

Abstract page for arXiv paper 2512.18809: FedVideoMAE: Efficient Privacy-Preserving Federated Video Moderation

arXiv - AI · 4 min ·
[2512.08980] Training Multi-Image Vision Agents via End2End Reinforcement Learning
Machine Learning

[2512.08980] Training Multi-Image Vision Agents via End2End Reinforcement Learning

Abstract page for arXiv paper 2512.08980: Training Multi-Image Vision Agents via End2End Reinforcement Learning

arXiv - AI · 4 min ·

All Content

[2602.20196] OpenPort Protocol: A Security Governance Specification for AI Agent Tool Access
Ai Safety

[2602.20196] OpenPort Protocol: A Security Governance Specification for AI Agent Tool Access

The OpenPort Protocol introduces a governance-first approach for AI agents, ensuring secure access to application tools while addressing ...

arXiv - AI · 4 min ·
[2602.20191] MoBiQuant: Mixture-of-Bits Quantization for Token-Adaptive Elastic LLMs
Llms

[2602.20191] MoBiQuant: Mixture-of-Bits Quantization for Token-Adaptive Elastic LLMs

The paper presents MoBiQuant, a novel quantization framework for elastic large language models (LLMs) that adapts weight precision based ...

arXiv - Machine Learning · 4 min ·
[2602.20169] Autonomous AI and Ownership Rules
Robotics

[2602.20169] Autonomous AI and Ownership Rules

This article explores the ownership rules surrounding AI-generated outputs, examining how they are linked to their creators and the impli...

arXiv - AI · 3 min ·
[2601.12815] Multimodal Multi-Agent Empowered Legal Judgment Prediction
Ai Infrastructure

[2601.12815] Multimodal Multi-Agent Empowered Legal Judgment Prediction

This paper presents JurisMMA, a novel framework for Legal Judgment Prediction (LJP) that utilizes multimodal data to enhance the accuracy...

arXiv - AI · 4 min ·
[2602.21061] Tool Building as a Path to "Superintelligence"
Llms

[2602.21061] Tool Building as a Path to "Superintelligence"

The paper explores how Large Language Models (LLMs) can achieve superintelligence through the Diligent Learner framework, emphasizing the...

arXiv - AI · 3 min ·
[2602.20934] Architecting AgentOS: From Token-Level Context to Emergent System-Level Intelligence
Llms

[2602.20934] Architecting AgentOS: From Token-Level Context to Emergent System-Level Intelligence

The paper introduces AgentOS, a conceptual framework that transitions Large Language Models from static inference engines to dynamic cogn...

arXiv - AI · 3 min ·
[2602.20732] CHESS: Context-aware Hierarchical Efficient Semantic Selection for Long-Context LLM Inference
Llms

[2602.20732] CHESS: Context-aware Hierarchical Efficient Semantic Selection for Long-Context LLM Inference

The paper presents CHESS, a novel KV-cache management system designed for long-context LLM inference, enhancing efficiency and throughput...

arXiv - AI · 3 min ·
[2602.20728] Balancing Multiple Objectives in Urban Traffic Control with Reinforcement Learning from AI Feedback
Llms

[2602.20728] Balancing Multiple Objectives in Urban Traffic Control with Reinforcement Learning from AI Feedback

This paper explores the use of reinforcement learning from AI feedback (RLAIF) to balance multiple objectives in urban traffic control, a...

arXiv - AI · 3 min ·
[2602.20710] Counterfactual Simulation Training for Chain-of-Thought Faithfulness
Llms

[2602.20710] Counterfactual Simulation Training for Chain-of-Thought Faithfulness

The paper introduces Counterfactual Simulation Training (CST), a method designed to enhance Chain-of-Thought (CoT) faithfulness in large ...

arXiv - AI · 4 min ·
[2602.20708] ICON: Indirect Prompt Injection Defense for Agents based on Inference-Time Correction
Llms

[2602.20708] ICON: Indirect Prompt Injection Defense for Agents based on Inference-Time Correction

The paper introduces ICON, a novel framework designed to defend Large Language Model (LLM) agents against Indirect Prompt Injection (IPI)...

arXiv - AI · 3 min ·
[2602.20659] Recursive Belief Vision Language Model
Llms

[2602.20659] Recursive Belief Vision Language Model

The Recursive Belief Vision Language Model (RB-VLA) addresses limitations in current vision-language-action models by introducing a belie...

arXiv - AI · 4 min ·
[2602.20628] When can we trust untrusted monitoring? A safety case sketch across collusion strategies
Machine Learning

[2602.20628] When can we trust untrusted monitoring? A safety case sketch across collusion strategies

This paper explores the challenges of ensuring safety in AI systems using untrusted monitoring. It develops a taxonomy of collusion strat...

arXiv - AI · 4 min ·
[2602.20571] CausalReasoningBenchmark: A Real-World Benchmark for Disentangled Evaluation of Causal Identification and Estimation
Machine Learning

[2602.20571] CausalReasoningBenchmark: A Real-World Benchmark for Disentangled Evaluation of Causal Identification and Estimation

The CausalReasoningBenchmark introduces a new framework for evaluating automated causal inference, distinguishing between identification ...

arXiv - AI · 4 min ·
[2602.20502] ActionEngine: From Reactive to Programmatic GUI Agents via State Machine Memory
Llms

[2602.20502] ActionEngine: From Reactive to Programmatic GUI Agents via State Machine Memory

The paper presents ActionEngine, a novel framework that enhances GUI agents by transitioning from reactive execution to programmatic plan...

arXiv - Machine Learning · 4 min ·
[2602.20333] DMCD: Semantic-Statistical Framework for Causal Discovery
Llms

[2602.20333] DMCD: Semantic-Statistical Framework for Causal Discovery

The DMCD framework integrates LLM-based semantic drafting with statistical validation for causal discovery, enhancing performance across ...

arXiv - AI · 3 min ·
Machine Learning

AI energy use: New tools show which model consumes the most power, and why

The article discusses new tools that analyze the energy consumption of various AI models, highlighting the importance of understanding po...

AI Events · 1 min ·
Pentagon gives AI firm ultimatum: lift military limits by Friday or lose $200M deal
Ai Safety

Pentagon gives AI firm ultimatum: lift military limits by Friday or lose $200M deal

The Pentagon has issued an ultimatum to AI firm Anthropic, demanding the removal of military use restrictions on its Claude AI by Friday ...

AI Tools & Products · 8 min ·
Nvidia challenger AI chip startup MatX raised $500M | TechCrunch
Ai Infrastructure

Nvidia challenger AI chip startup MatX raised $500M | TechCrunch

MatX, an AI chip startup founded by ex-Google engineers, has raised $500M in Series B funding to develop processors aimed at outperformin...

TechCrunch - AI · 4 min ·
AI Ready? Charting a Shared Course for Adoption
Ai Infrastructure

AI Ready? Charting a Shared Course for Adoption

The article discusses the rapid integration of AI into Indiana's economy, emphasizing the need for collaboration among business, educatio...

AI News - General · 4 min ·
Innovation on the move | MIT Technology Review
Ai Infrastructure

Innovation on the move | MIT Technology Review

MIT alumni are transforming the MBTA by enhancing route planning, improving service, and fostering a culture of innovation to better conn...

MIT Technology Review · 17 min ·
Previous Page 88 Next

Related Topics

Stay updated with AI News

Get the latest news, tools, and insights delivered to your inbox.

Daily or weekly digest • Unsubscribe anytime