Research & Benchmarks - Page 4 of 28
Academic AI research, performance benchmarks, scientific breakthroughs, and peer-reviewed studies advancing artificial intelligence frontiers.
547 articles
View complete article list
LLM pipeline compares DAO ERC‑8004 and Google A2A governance, 4,323 records
Everyone building AI governance has a favorite story. One side says permissionless systems breed chaos. The other says corporate models centralize control. Both stories are wrong, and now there's data to prove it.
Calibration uses NVIDIA Triton Llama-3-8B A10 and vLLM Qwen2.5-7B RTX 4090 data
Inference is a story of two systems, and the story begins with a single millisecond. Consider a transaction authorization. The clock starts with the ISO 8583 budget.
Figma launches AI motion graphics, shader tools, code layers, and new creative materials
Figma's latest update targets more than interface clutter. It takes aim at the org chart. At its Config conference, the company baked motion graphics, shader effects, and live code layers directly into its design canvas. You prompt.
NVIDIA RTX PRO 4500 Blackwell GPUs Power New Amazon EC2 G7 Instances
Amazon just dropped new chips into its cloud. They’re banking you’ll pay for them. The EC2 G7 instances are now powered by NVIDIA’s RTX PRO 4500 Blackwell GPUs. This is a direct upgrade to the previous G6 generation. The numbers are stark.
Stanford researchers present agentic AI 'scientists' at VB Transform 2026
Forget the lone genius in the lab. The next big discovery might be managed by a CEO made of code. Stanford researchers are pitching this at VB Transform 2026. They've built a hierarchy of AI agents to act as a full scientific team.
Turning Logistic Regression Coefficients into Credit Score Grid
Logistic regression spits out coefficients, not a usable credit score. To build one, start with a single, concrete fact: for any variable, the category with the strongest positive link to default gets a baseline of zero points.
Metric-Dependent Annotation Saturation for Learning from Label Distributions
We treat annotator disagreement like static to filter out. Maybe we're filtering out the point. A new study asks how many people you actually need to pin that signal down. The answer isn't one number. It depends entirely on what you're measuring.
DFlash speculative decoding boosts NVIDIA Blackwell inference up to 15×
NVIDIA Blackwell delivers 15 petaflops of dense NVFP4 compute, a staggering amount of raw power. But raw power means nothing if the pipeline can’t feed it. That’s where DFlash comes in.
NVIDIA BioNeMo Toolkit Enables AI Scientist to Align, Fold, and Dock Molecules
Lab work is slow. The AI scientist is not. It runs without sleep or salary, folding proteins and aligning sequences in a silent digital loop. NVIDIA’s BioNeMo Toolkit is building it a new lab.
Data2Story converts CSVs to articles with 7 AI; 53 readers prefer them to human
The numbers are stark. Fifty-three readers, when given a choice, preferred the machine. Not just any machine, a seven-agent AI system called Data2Story that ingests a CSV and spits out a full, interactive news article.
M* introduces overlapped scheduling to streamline multimodal model serving
Multimodal models are slow. The real problem isn't the math. It's the wait. The GPU sits idle while the CPU schedules the next batch, a tax paid on every loop. M* cuts that tax.
Benchmark shows Claude Fable 5 passes only 3% of tasks, 31 of 91 fail 50%
Claude Fable 5 achieved perfection exactly three times. That’s the stark finding from a new benchmark testing 91 real-world tasks, where the top model’s flawless scorecard reads a paltry 3%.
Nobel laureate John Jumper departs DeepMind for Anthropic after AlphaFold win
Winning a Nobel Prize buys you a golden ticket. John Jumper just cashed his in at Anthropic’s door.
OpenAI shows small 'beneficial trait' training makes AI safer, less manipulable
Engineers at OpenAI have a new trick for making AI systems less easily corrupted. It’s a small but consequential tweak to the final stage of training.
Google DeepMind uses MITRE ATT&CK to monitor AI agents as rogue employees
Google DeepMind has a new security headache, and its source isn't human. The unit is now surveilling its own advanced AI agents as potential insider threats, applying a security framework built to catch sophisticated human hackers.
DeFAb Benchmark Enforces Polynomial-Time Checks for Logical Rigor
Reasoning is the bottleneck. Not prose, not parameters, but the quiet, unforgiving machinery of logical structure. DeFAb, a new benchmark, directly weaponizes polynomial-time verifiability against that bottleneck.
OpenAI researchers aim to forecast AI model failure rates pre‑launch
Testing AI models before launch is like stress-testing a car on a closed track: revealing, but rarely a mirror of real highways. Researchers fabricate adversarial prompts, hunt for known failure modes, and declare the model ready.
Nvidia AI Agent Trains Robots Autonomously, Editing Code from Papers
Nvidia has built a robot that reads academic papers and rewrites its own training code. It corrects itself. The system, which the company calls an AI agent, works in two parts. The first part needs a person. The second does not.
XGBoost, ALBERT, BioBERT, Med‑LLaMA evaluated for pharmacovigilance
Pharmacovigilance is not a game of scale. It is a game of precision. When millions of adverse event reports flood databases, the task of linking a drug to a harm demands more than raw parameter counts, it demands models that understand clinical...
OpenAI's Deployment Simulation Beats Baseline, Adds Risk Checks to Agentic Code
The next generation of AI doesn’t just write code, it acts on it. That shift brings a new class of risk: models that behave differently when they suspect they’re being watched. OpenAI’s latest research surfaces a remedy.
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.