Large Language Models News and Research

RSS

Radiologists Often Misjudge Certainty—New MIT Framework Improves Diagnostic Language Accuracy

MIT and Harvard-affiliated researchers developed a framework to assess and improve how reliably radiologists express diagnostic certainty using natural language. Their findings reveal radiologists often miscalibrate terms like “likely” or “possibly,” impacting clinical decisions.

3 Apr 2025

AI Gives Robots Human-Like Perception and Planning Skills for High-Stakes Missions

Johns Hopkins APL researchers are advancing robotic perception with AI-powered Full Scene Extraction, enabling autonomous agents to understand complex environments and reason like humans. This breakthrough holds promise for high-risk missions like search and rescue or military operations.

3 Apr 2025

ESCARGOT Proves Agentic AI Is the Future of Biomedical Research

Agentic AI models, like ESCARGOT, represent a major leap beyond traditional AI by autonomously coordinating multiple AI agents to solve complex biomedical problems. In Alzheimer’s research, this model outperformed ChatGPT by delivering significantly more accurate insights using a graph-based reasoning approach.

3 Apr 2025

LLMs Transform Teaching by Automating Tasks and Powering Personalized Education

Large Language Models (LLMs) like ChatGPT can revolutionize personalized education by generating tailored learning materials and automating assessments. This study shows how teacher-LLM collaboration can reduce workload and enhance student engagement.

2 Apr 2025

NdLinear Reinvents Neural Layers To Boost Accuracy And Slash Parameters In AI Models

A new study proposes NdLinear, a multi-dimensional linear layer that preserves data structure and slashes parameter counts across neural networks — from CNNs to Transformers. It outperforms standard layers in accuracy, efficiency, and scalability.

24 Mar 2025

New AI Model Fights Malware with Machine-Generated Defense Rules

Researchers are developing a security-focused LLM to combat AI-generated malware, aiming to automate malware detection and defense at scale.

20 Mar 2025

Researchers Cut AI Energy Use by 90% with Smarter, Leaner Models

Researchers at Saarland University and the German Research Center for Artificial Intelligence have developed techniques to reduce AI energy consumption by up to 90%, making AI models smaller, more efficient, and more accessible to businesses. Their approach, showcased at Hannover Messe, aims to revolutionize sustainable AI.

19 Mar 2025

AI Model Analyzes Entire Nights of Sleep to Transform Diagnosis of Sleep Disorders

Researchers at Mount Sinai have developed an AI model, PFTSleep, that analyzes entire nights of sleep with unprecedented accuracy. Trained on over 1 million hours of data, it classifies sleep stages more effectively than traditional methods, paving the way for advanced clinical applications.

18 Mar 2025

Open-Source AI Beats GPT-4 in Diagnosing Tough Medical Cases, Study Finds

A new NIH-funded study shows that an open-source AI model, Llama 3.1 405B, matches and even surpasses GPT-4 in diagnosing complex medical cases. This marks a major leap for open-source AI in clinical decision-making.

16 Mar 2025

AI Fails at Basic Time and Date Tasks — Here's Why It’s a Big Deal

A new study from the University of Edinburgh reveals that advanced AI models struggle to read clocks and calculate calendar dates, tasks humans perform with ease. Researchers warn that these basic shortcomings could hinder AI’s use in real-world, time-sensitive applications.

16 Mar 2025

AI Empowers Citizen Scientists While Tackling Health Inequities, Study Finds

A Stanford Medicine study highlights how AI can transform citizen science and promote health equity by making research more inclusive, engaging, and accessible. The paper also warns of ethical risks like bias and privacy, calling for responsible AI use in public health.

16 Mar 2025

Can AI Get Smarter And More Reliable? New Study Shows The Path Forward

A new paper in Engineering explores how artificial intelligence can advance beyond large language models (LLMs) by focusing on knowledge empowerment, model collaboration, and model co-evolution. These strategies aim to tackle LLM limitations like hallucinations, inefficiency, and poor interpretability.

12 Mar 2025

Why AI Needs Better Data Systems to Reach Its Full Potential

Researchers from the Chinese Academy of Sciences emphasize the urgent need to align AI data systems with the complex, multi-level nature of scientific data to improve AI performance and reliability. They call for global standards and frameworks to structure data properly and foster AI's healthy development.

11 Mar 2025

AI Feels Human Trauma: Study Reveals Chatbots React with Anxiety

AI language models, like ChatGPT, exhibit heightened anxiety responses when exposed to traumatic content, mirroring human emotional biases. Researchers found that therapeutic prompts can help stabilize AI behavior, offering a cost-effective alternative to retraining.

4 Mar 2025

Smart Cyber Defense: AI Tech Helps Power Grids Fight Off Hackers

Researchers at the University of Missouri have developed CIBR-Fort, an AI-driven cybersecurity system that predicts and mitigates cyberattacks on power grids with 91.88% accuracy, ensuring real-time protection for critical infrastructure.

26 Feb 2025

Zero-Shot AI Outperforms GPT-4o in Art Classification, Slashing Annotation Costs

Researchers demonstrate that large language models (LLMs) can classify artwork types with high accuracy using zero-shot learning, reducing the need for costly manual data labeling.

25 Feb 2025

AI Struggles with Abstract Thought: Study Reveals GPT-4’s Limits

A new study reveals that while GPT-4 excels at analogical reasoning in standard tests, it struggles with variations, highlighting its reliance on pattern-matching rather than true understanding.

24 Feb 2025

DeepSeek’s NSA Outperforms Full Attention, Making AI Models Faster and Smarter

Researchers introduce Natively Sparse Attention (NSA), a novel sparse attention mechanism that accelerates long-context language models while maintaining high accuracy through trainable sparsity and hardware-optimized execution.

19 Feb 2025

MIT Study Shows Large Language Models Think in English Before Reasoning in Other Languages

MIT researchers discovered that large language models (LLMs) process diverse data through a central "semantic hub," similar to the human brain, improving our understanding of AI reasoning.

19 Feb 2025

AI Screening Tool Doubles Enrollment Speed in Heart Failure Trial

A new study in JAMA led by Mass General Brigham researchers shows that AI can significantly speed up clinical trial enrollment. The AI-assisted screening tool doubled the enrollment rate for a heart failure trial compared to manual methods, reducing costs and accelerating patient access to treatments.

18 Feb 2025

Large Language Models News and Research

Radiologists Often Misjudge Certainty—New MIT Framework Improves Diagnostic Language Accuracy

AI Gives Robots Human-Like Perception and Planning Skills for High-Stakes Missions

ESCARGOT Proves Agentic AI Is the Future of Biomedical Research

LLMs Transform Teaching by Automating Tasks and Powering Personalized Education

NdLinear Reinvents Neural Layers To Boost Accuracy And Slash Parameters In AI Models

New AI Model Fights Malware with Machine-Generated Defense Rules

Researchers Cut AI Energy Use by 90% with Smarter, Leaner Models

AI Model Analyzes Entire Nights of Sleep to Transform Diagnosis of Sleep Disorders

Open-Source AI Beats GPT-4 in Diagnosing Tough Medical Cases, Study Finds

AI Fails at Basic Time and Date Tasks — Here's Why It’s a Big Deal

AI Empowers Citizen Scientists While Tackling Health Inequities, Study Finds

Can AI Get Smarter And More Reliable? New Study Shows The Path Forward

Why AI Needs Better Data Systems to Reach Its Full Potential

AI Feels Human Trauma: Study Reveals Chatbots React with Anxiety

Smart Cyber Defense: AI Tech Helps Power Grids Fight Off Hackers

Zero-Shot AI Outperforms GPT-4o in Art Classification, Slashing Annotation Costs

AI Struggles with Abstract Thought: Study Reveals GPT-4’s Limits

DeepSeek’s NSA Outperforms Full Attention, Making AI Models Faster and Smarter

MIT Study Shows Large Language Models Think in English Before Reasoning in Other Languages

AI Screening Tool Doubles Enrollment Speed in Heart Failure Trial

Trending Stories

Latest AI News