Research
Chinese Team Uses Quantum Entanglement to Block Hardware Backdoors
A breakthrough in quantum data transmission by a Chinese research team could end hardware-based cyberattacks. Is this the next frontier in secure communication?
Plasmon Computing Could Slash AI’s Power Demand
Hector De Los Santos and his team are pioneering plasmon computing—using electron waves instead of current—to tackle AI’s growing energy problem.
DeepRead: RAG Agent Uses Document Structure to Deliver Smarter Answers
Researchers unveil DeepRead, a new Retrieval-Augmented Generation (RAG) agent that reads document structure to boost question-answering accuracy.
RaBiT Sets New Standard for 2-Bit LLM Quantization
RaBiT, a new quantization method, solves feature co-adaptation in residual binarization, delivering state-of-the-art 2-bit performance and major inference speed gains for large language models.
New AI Model Integrates Multimodal Data to Improve Ameloblastoma Diagnosis
Researchers introduce an AI model and dataset combining radiology, pathology, and clinical data to enhance diagnosis and surgical planning for ameloblastoma.
Study Finds LLM Sequence Lengths That Cut Energy Use on NVIDIA H100 GPUs
New research shows optimizing input and output lengths can slash energy consumption during large language model inference.
ProAct Framework Boosts Long-Horizon Planning in LLM Agents
New framework combines Grounded LookAhead Distillation and Monte-Carlo Critic to improve planning accuracy, matching closed-source models.
New Framework Aims to Tame LLM Hallucinations in Security Management
Researchers propose an iterative loop with consistency checks and digital twins to improve the reliability of LLMs for incident response, potentially cutting recovery times by 30%.
Study Finds Bias in Leading LLMs Assessing Financial Risk
New research exposes bias in top LLMs like GPT-5.1 and Gemini-2.5 Pro when judging financial risk, highlighting risks for real-world deployment.
OpenAI Invests in Merge Labs to Advance Brain-Computer Interfaces
OpenAI’s new investment signals a push to merge biological and artificial intelligence, sparking excitement and ethical debate over enhancing human abilities.
New 'recos' Metric Aims to Redefine Semantic Similarity
A novel similarity metric called 'recos' is challenging the long-standing dominance of cosine similarity in semantic analysis by better capturing nonlinear relationships.
AI Hunts for New Physics as Particle Physics Faces Crisis
With the Large Hadron Collider not yielding expected insights, physicists are turning to unsupervised machine learning to uncover new patterns and probe for new physics.
Facebook Research Unveils JEPA-WMs: A Leap Forward in AI Planning
New model optimizes planning in learned representation spaces, outperforming established baselines in navigation and manipulation tasks, paving the way for more efficient AI agents.
MLLMs Flounder in Spatial Reasoning, New VLN-MME Benchmark Reveals
A new evaluation framework, VLN-MME, shows Multimodal Large Language Models struggle with context and spatial reasoning in Vision-and-Language Navigation tasks, even with Chain-of-Thought prompting.
New Framework Aims to Bridge the Gap in Multi-Modal AI for Autonomous Systems
Researchers propose a unified approach to pre-training AI models using diverse sensor data, tackling the challenges of spatial intelligence in autonomous systems.