AI Pulse News

Houthi threat spikes US-Iran ceasefire odds as Bab el-Mandeb tensions rise

AVAX One pivots to “AI infra + Bitcoin hash” with 10MW Alberta microgrid

Cysic’s Venus zkVM goes open source as Ethereum eyes proof markets

Eightco turns into a listed bet on Worldcoin and OpenAI

Proposed Ethereum Standard Aims to Help AI Agents Execute Complex DeFi Trades

ASML (ASML) Stock Tumbles 4% After US Lawmakers Propose China Export Ban

Xiao-I (AIXI) Stock Explodes Over 500% Following Major Chinese Supreme Court Victory Against Apple

Meta to roll out open source AI models in next phase of development

Anthropic’s refusal to arm AI is exactly why the UK wants it

Broadcom (AVGO) Surges on Multi-Billion Dollar AI Chip Partnerships with Google and Anthropic

Intel is going all-in on advanced chip packaging

JPMorgan CEO says AI will transform banking faster than the internet era

Khamenei unconscious in Qom raises concerns over Iran’s leadership stability

Bitcoin miners face a new rival for cheap power as Anthropic signs multi-gigawatt compute deal

Position: Science of AI Evaluation Requires Item-level Benchmark Data

Researchers argue that current AI evaluation methods have systemic validity failures and propose item-level benchmark data as essential for rigorous AI evaluation. They introduce OpenEval, a repository of item-level benchmark data to support evidence-centered AI evaluation and enable fine-grained diagnostic analysis.

Toward Full Autonomous Laboratory Instrumentation Control with Large Language Models

Researchers demonstrate how large language models like ChatGPT can automate laboratory instrument control, reducing programming barriers for scientists. The study shows LLMs can create custom scripts and operate as autonomous AI agents for lab equipment management.

VERT: Reliable LLM Judges for Radiology Report Evaluation

Researchers introduced VERT, a new LLM-based metric for evaluating radiology reports that shows up to 11.7% better correlation with radiologist judgments compared to existing methods. The study demonstrates that fine-tuned smaller models can achieve significant performance gains while reducing inference time by up to 37.2 times.

When Adaptive Rewards Hurt: Causal Probing and the Switching-Stability Dilemma in LLM-Guided LEO Satellite Scheduling

Research reveals that adaptive reward mechanisms in AI-guided satellite scheduling systems actually hurt performance, with static reward weights achieving 342.1 Mbps versus dynamic weights at only 103.3 Mbps. The study found that fine-tuned LLMs performed poorly due to weight oscillation issues, while simpler MLP models achieved superior results of 357.9 Mbps.

Selective Forgetting for Large Reasoning Models

Researchers propose a new framework for 'selective forgetting' in Large Reasoning Models (LRMs) that can remove sensitive information from AI training data while preserving general reasoning capabilities. The method uses retrieval-augmented generation to identify and replace problematic reasoning segments with benign placeholders, addressing privacy and copyright concerns in AI systems.

Rashomon Memory: Towards Argumentation-Driven Retrieval for Multi-Perspective Agent Memory

Researchers propose Rashomon Memory, a new AI agent memory architecture where multiple goal-conditioned agents maintain parallel interpretations of the same events and negotiate through argumentation at query time. The system allows AI agents to handle conflicting perspectives on experiences rather than forcing a single interpretation, using Dung's argumentation semantics to determine which proposals survive retrieval.

PRAISE: Prefix-Based Rollout Reuse in Agentic Search Training

Researchers introduce PRAISE, a new framework that improves training efficiency for AI agents performing complex search tasks like multi-hop question answering. The method addresses key limitations in current reinforcement learning approaches by reusing partial search trajectories and providing intermediate rewards rather than only final answer feedback.

Structured Multi-Criteria Evaluation of Large Language Models with Fuzzy Analytic Hierarchy Process and DualJudge

Researchers developed DualJudge, a new framework for evaluating large language models that combines structured Fuzzy Analytic Hierarchy Process (FAHP) with traditional direct scoring methods. The approach addresses inconsistent LLM evaluation by incorporating uncertainty-aware reasoning and achieved state-of-the-art performance on JudgeBench testing.

FactReview: Evidence-Grounded Reviews with Literature Positioning and Execution-Based Claim Verification

Researchers introduce FactReview, an AI system that improves academic peer review by combining claim extraction, literature positioning, and code execution to verify research claims. The system addresses weaknesses in current LLM-based reviewing by grounding assessments in external evidence rather than relying solely on manuscript narratives.

Compliance-by-Construction Argument Graphs: Using Generative AI to Produce Evidence-Linked Formal Arguments for Certification-Grade Accountability

Researchers propose a compliance-by-construction architecture that integrates Generative AI with structured formal argument representations to ensure accountability in high-stakes decision systems. The approach uses typed Argument Graphs, retrieval-augmented generation, validation constraints, and provenance ledgers to prevent AI hallucinations while maintaining traceability for regulatory compliance.

Profile-Then-Reason: Bounded Semantic Complexity for Tool-Augmented Language Agents

Researchers introduce Profile-Then-Reason (PTR), a new framework for AI language agents that use external tools, which reduces computational overhead by pre-planning workflows rather than recomputing after each step. The approach limits language model calls to 2-3 times maximum and shows superior performance in 16 of 24 test configurations compared to reactive execution methods.