Multimodal AI
Researchers introduce MMR-Bench, a comprehensive benchmark evaluating how well routing systems direct queries to optimal multimodal LLMs across diverse visual reasoning tasks.
AI Safety
New research reveals critical gaps in how human experts evaluate AI safety in mental health applications, questioning whether current testing methods can reliably identify harmful model behaviors.
synthetic data
New survey introduces systematic metrics for evaluating synthetic data quality and trustworthiness from LLMs, addressing critical challenges in detecting and assessing AI-generated content reliability.
LLM Research
New research combines graph-based local reasoning with belief propagation to help LLMs tackle complex investigative tasks, enabling more reliable multi-step analysis in AI systems.
LLM Research
New research challenges how we measure edit locality in LLM model editing, revealing potential blind spots in current evaluation methods that could impact knowledge modification reliability.
neuro-symbolic AI
New research unifies Datalog symbolic reasoning with neural computation via tensor contractions, enabling differentiable logic programming with potential implications for AI reasoning systems.
Reinforcement Learning
LinkedIn shares practical insights from training agentic reinforcement learning systems for GPT-OSS, covering infrastructure challenges, reward modeling, and lessons learned from real-world deployment.
Alibaba
Alibaba unveils Qwen3-Max-Thinking, a reasoning-focused AI model that outperforms rivals in select benchmarks, intensifying competition in the large language model space.
Microsoft
Microsoft announces its Maia 200 custom AI accelerator, entering direct competition with Amazon and Google in the race to build proprietary silicon for AI workloads.
Multi-Agent Systems
AgentScope provides a flexible framework for orchestrating multiple LLM agents with built-in communication protocols, fault tolerance, and scalability features for complex AI workflows.
Synthesia
AI avatar startup Synthesia hits $4 billion valuation with secondary share sale, allowing employees to cash out as synthetic video market matures.
Neural Networks
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.