·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Structured Authority Injection: Persistent Human Control for LLM Agents via Typed MCP Overrides36m◆Rethinking User Encoders: Routing Diffusion States for Scenario-Adaptive User Representation36m◆From Events to Reasoning: Hierarchical Rule Composition over Explicit States36m◆ECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation36m◆Sensitivity Analysis of Decision Rules in an Agent-Based Agro-Ecological Model36m◆PRIME: Towards Verifiable Evidence-Grounded Reasoning in LLM Agents36m◆From Natural Language Intent to Executable Control: Hierarchical Feedback-Guided LLM Planning for Multi-component Cyber-Physical Systems36m◆Null Controls Are Necessary but Not Sufficient: Evaluating Acceptance Guards for Scientific Discovery and Interpretability36m◆Learnable Splines for Unstructured Neural Operators36m◆Functional Welfare for AI Safety36m◆Author-in-the-loop Agentic Reproducibility Auditing for a Machine Learning Conference36m◆Harness-Delta Attribution: Dissecting Gains From Harness Evolution36m◆Testing the Planning Claims of LLM-Based Agents against Formal Semantics36m◆Are Large Reasoning Models Good Everyday Reasoners? A Comparison with Non-Reasoning Models on Logical Reasoning Tasks36m◆FRACTAL: A Self-Similar Multi-Agent Orchestration Architecture with Formal Communication, Convergence, and Depth Bounds36m◆Repairing LLM-Generated Specifications with Deterministic Diagnosis36m◆At the Limits of Computability: Steps Towards Autonomous AI Discovery on the Busy Beaver Frontier36m◆From Relevant Memories to Sufficient Evidence: Rethinking Memory Utilization in LLM Agents36m◆Not All Graphs Are Maps: A Task-Agnostic Exploration Framework for Mobile GUI Navigation36m◆FlexUMM: Flexible Representation Learning for Unified Multimodal Models36m◆Structured Authority Injection: Persistent Human Control for LLM Agents via Typed MCP Overrides36m◆Rethinking User Encoders: Routing Diffusion States for Scenario-Adaptive User Representation36m◆From Events to Reasoning: Hierarchical Rule Composition over Explicit States36m◆ECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation36m◆Sensitivity Analysis of Decision Rules in an Agent-Based Agro-Ecological Model36m◆PRIME: Towards Verifiable Evidence-Grounded Reasoning in LLM Agents36m◆From Natural Language Intent to Executable Control: Hierarchical Feedback-Guided LLM Planning for Multi-component Cyber-Physical Systems36m◆Null Controls Are Necessary but Not Sufficient: Evaluating Acceptance Guards for Scientific Discovery and Interpretability36m◆Learnable Splines for Unstructured Neural Operators36m◆Functional Welfare for AI Safety36m◆Author-in-the-loop Agentic Reproducibility Auditing for a Machine Learning Conference36m◆Harness-Delta Attribution: Dissecting Gains From Harness Evolution36m◆Testing the Planning Claims of LLM-Based Agents against Formal Semantics36m◆Are Large Reasoning Models Good Everyday Reasoners? A Comparison with Non-Reasoning Models on Logical Reasoning Tasks36m◆FRACTAL: A Self-Similar Multi-Agent Orchestration Architecture with Formal Communication, Convergence, and Depth Bounds36m◆Repairing LLM-Generated Specifications with Deterministic Diagnosis36m◆At the Limits of Computability: Steps Towards Autonomous AI Discovery on the Busy Beaver Frontier36m◆From Relevant Memories to Sufficient Evidence: Rethinking Memory Utilization in LLM Agents36m◆Not All Graphs Are Maps: A Task-Agnostic Exploration Framework for Mobile GUI Navigation36m◆FlexUMM: Flexible Representation Learning for Unified Multimodal Models36m◆
News/HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding
arxiv
PublishedMay 8, 2026 at 4:00 AM
—neutral

HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2601.14724v4 Announce Type: replace-cross Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated significant improvement in offline video understanding. However, extending these capabilities to streaming video inputs, remains challenging, as existing models

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivStructured Authority Injection: Persistent Human Control for LLM Agents via Typed MCP Overrides36marxivRethinking User Encoders: Routing Diffusion States for Scenario-Adaptive User Representation36marxivFrom Events to Reasoning: Hierarchical Rule Composition over Explicit States36marxivECHOES: Seeing Beyond the Current Image via Recursive History Compression and Privileged On-Policy Distillation36m
The Bubble Brief
MON · WED · FRI

Read AI insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews