Research & Benchmarks - Page 3 of 24
Academic AI research, performance benchmarks, scientific breakthroughs, and peer-reviewed studies advancing artificial intelligence frontiers.
474 articles
View complete article list
Explainable ML Classifies Alzheimer's Early in 1,641 ADNI Subjects
Alzheimer’s disease touches more than 55 million people worldwide, yet clinicians still lack a reliable, interpretable way to separate normal cognition, mild cognitive impairment and full‑blown dementia using everyday assessments.
MIT researchers train AI to read charts, streamlining downstream workflows
Enterprises chasing faster decisions often lean on generative AI to sift through the sea of charts in market briefs and financial reports.
Lightweight CNN Boosts Adversarial Robustness in EEG‑Based Brain‑Computer Interfaces
Brain‑computer interfaces that read EEG signals have become more accurate, but their safety has lagged behind. Why does that matter?
Hundreds sign Leiden Declaration as AI threatens mathematicians' profession
Mathematicians are sounding the alarm. A declaration drafted by a working group of 16 researchers over eight months—after a September 2025 conference at Leiden University—has landed on the International Mathematical Union’s desk and, as of June 2,...
Transformer tops Gait2Hip-60 benchmark with 0.819 R² in hip force prediction
Why does this matter? Predicting hip muscle forces and joint moments during walking has long depended on musculoskeletal simulation—accurate but slow, and hard to bring into a clinic.
QASM-Eval Introduces First Dataset for Training LLMs on OpenQASM-3
Why does this matter? Because quantum computing is still stuck in the noisy‑intermediate‑scale‑quantum (NISQ) era, where every ounce of hardware control counts.
Parallax adds learned covariance correction to linear attention, retains softmax
The Transformer’s attention has barely moved since 2017, and most efficiency research has tried to toss softmax out entirely. Here’s the thing: a new paper decides to keep softmax and simply bolt a correction branch onto it.
Men use AI coding agents over twice as often as women; economists at 39%
Anthropic’s latest survey shines a light on how social scientists are adopting AI‑driven coding assistants.
Molecule-trained AI gives better chicken pairing suggestions than recipe AI
The startup Kaikaku.AI is putting a spotlight on how an AI’s training data shapes the food pairings it suggests.
AI search agents favor confirming hits, sideline gut answers, study finds
Why does this matter? Because the promise of AI‑driven search agents has always been that they can crawl the web, stitch together fresh facts, and hand you an answer you couldn’t have guessed.
OpenAI gives free life‑sciences AI model to aid government pandemic prep
OpenAI is rolling out a new initiative called the Rosalind Biodefense program, offering free access to its life‑sciences AI model, GPT‑Rosalind.
Review paper claims code defines AI agents' reasoning and behavior
A new review paper co‑authored by researchers at the University of Illinois Urbana‑Champaign, Meta, and Stanford puts code front and centre in the conversation about AI agents.
NVIDIA research moves robotics simulation to reality, revealing robot confusion
Why does this matter? Robots still stumble when the world is messy. In a demo on the PEEK project page, a robot is asked to “give the banana to NVIDIA founder and CEO Jensen Huang.” The table holds a photo of Huang, a picture of Michael Jordan, a...
CVPR 2026 Friday Session: STARFlow‑V Video Modeling Poster #178, 4‑6 PM
Apple is back at the IEEE/CVF Conference on Computer Vision and Pattern Recognition, taking place in person at Denver’s Colorado Convention Center from June 3‑7.
USD E^3USD ‑Agent splits fast router from LLM meta‑controller for edge inference
Edge deployments of generative AI are running into two practical headaches. First, the performance of each model on each device often isn’t known when the system is rolled out.
Sakana AI's DiffusionBlocks Apply Uniform [4,4,4] Layers Across Three Blocks
Why does training deep neural nets still choke on memory? Researchers at Sakana AI and the University of Tokyo think they’ve found a practical answer.
AI Agent Auto-Identifies Unreadable Model Parameters from CSV Files
The promise of AI‑driven optimization has been humming in the background of business decisions for years.
Learn to Build AI Projects: n8n Automation, Financial Data, Summaries, Reports
AI isn’t interesting because it looks cool on a demo screen; it matters when it takes the grunt work out of everyday tasks.
AI Agents Falter in Production as Backward Design Overburdens Model
When we finally dug into a stubborn failure, it took two days of debugging to see what was really happening. The model wasn’t hallucinating; the input‑output tools returned the right answers.
Hugging Face releases LeRobot Humanoid: 3D‑printable legs for robot research
A $2,500 pair of 3‑D‑printed legs is now available for anyone who wants to put AI‑driven software into a real‑world robot.
📚 Featured Resources & Reviews
🎓
No Code MBA Course Review
Learn to build AI-powered apps without coding. Our comprehensive review of No Code MBA's course.
🔧
AI Tools & Resources
Curated collection of AI tools, courses, and frameworks to accelerate your AI journey.
📬
Weekly AI Digest
Get the week's most important AI news delivered to your inbox every week.