Research & Benchmarks - Page 10 of 28
Academic AI research, performance benchmarks, scientific breakthroughs, and peer-reviewed studies advancing artificial intelligence frontiers.
547 articles
View complete article list
iTARFlow Shows Competitive Performance on ImageNet 64‑256px Resolutions
For years, normalizing flows were the quiet kids in the image generation class. Diffusion models and autoregressive transformers got all the attention, especially when dealing with bigger pictures.
CreativityBench benchmark introduces 4K‑entity affordance KB to test LLM creativity
The world isn't made of hammers and nails. It's a jumble of whatever's heavy enough to pound a nail when the real tool is missing. That simple, messy truth of human ingenuity completely baffles our best AI.
Self-Attentive Meta-Optimizer Adds Gradient Alignment and Group-Adaptive Rates
We’ve settled for dumb optimizers. The ubiquitous AdamW doesn't think; it just applies the same rule to every parameter update with the mechanical faith of a metronome. But gradients fight each other. Different tasks need different things.
Local edits in LLM-driven NAS can trigger broader performance shifts
Changing one tiny part of a neural network is supposed to be safe. A minor adjustment. Like swapping a transistor. In reality, it’s closer to pulling one thread in a sweater and watching the whole sleeve unravel.
Groq‑Powered Agentic Assistant Uses Sub‑Agent to Catalog 2024‑25 SLMs
The pace of small language model releases has become dizzying, scores of new architectures, benchmarks, and use‑cases flood the ecosystem every quarter.
AI autoencoders and joint communications‑sensing rank among 6G enablers
Forget a stronger signal. The engineers drafting the 6G standard at the IEEE are sketching something else entirely: a network that sees, thinks, and learns. Their first move is brutal. They're ripping out racks of dedicated hardware.
Anthropic adds 'dreaming' feature to Claude Managed Agents for memory recall
Anthropic just gave its corporate AI a feature named after sleep. "Dreaming." It's a slick bit of branding for what is, fundamentally, a scheduled background process.
Anthropic's USD 200 B, five‑year Google Cloud deal makes up >40% of backlog
The AI boom is measured in zeros and it's mostly a debt two companies can't afford. Anthropic just promised Google Cloud $200 billion over five years. Nobody announced that figure.
MRC retires paths, then probes to confirm failures and recovery
Packet loss plagues networks. Most guess at its cause. OpenAI's MRC system does not guess. When it detects loss, it immediately takes a path out of service. Then it runs a test. It must confirm a real failure and monitor for recovery.
eOptShrinkQ enables near‑lossless KV cache compression with spectral denoising
Compressing the key-value cache without destroying signal fidelity has long been a tug-of-war between quantization error and the noise that quantization itself introduces.
Harvard study finds OpenAI's o1 and 4o outdiagnose ER doctors in 76‑patient test
Seventy-six patients walked into a Boston emergency room. Two attending physicians examined them, made their calls. So did two AI models from OpenAI, o1 and 4o.
2021 EDEN-unbiased quantizer beats 2026 successor in average accuracy
A 2021 design is outperforming its 2026 successor , across every tested configuration, and by a wide margin.
US benchmark shows China lagging; Deepseek model underperforms private tests
The US government has declared China's best AI model is falling behind. It's a political verdict. DeepSeek V4 is the current Chinese champion.
Google DeepMind AI co‑clinician beats GPT‑5.4 in blind tests, lags docs
Medical AI has one job: get the answer right. A new system from Google DeepMind mostly does, which makes its single, critical failure that much more important.
Anthropic benchmark says Claude matches experts, 23 tasks remain ambiguous
Anthropic asked a new version of its Claude model to solve 99 problems in bioinformatics. On 76 tasks that at least one human expert could handle, Claude matched their performance. On the 23 that stumped every expert, it got about a third right.
Grok Voice Think Fast 1.0 lets non‑programmers design agents via console.x.ai
The barrier between imagination and execution just collapsed. You don’t need a single line of code to breathe life into a voice agent, console.x.ai’s playground gives you a blank slate and two paths to build.
WPI professor Gerych offers solution to AI vision ‘Whac‑a‑mole’ bias dilemma
Artificial intelligence vision systems are plagued by a stubborn paradox: scrub one bias, and another pops up elsewhere. This relentless game of algorithmic Whac-a-Mole has frustrated researchers for years.
New method advances privacy‑preserving AI training on consumer devices
The pitch was simple: keep your life locked on your phone. For on-device AI, that's the whole sell. Privacy isn't a bonus; it's the box. But making that promise work? A nightmare. Weak processors choke. Spotty signals drop.
Musk says he was duped, warns AI could kill us, xAI to IPO via SpaceX in June
Elon Musk built an empire selling the future. Now he’s in court selling a story about being its victim.
NVIDIA BioNeMo wraps CPU layer with DistributedTriangleMultiplication
The bottleneck has always been memory. In structural biology, protein complexes sprawl across thousands of residues, and single‑GPU limits have kept those models caged.
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.