Research news · Page 25

Best of AI, daily

Tue21

6:02 PMResearchReach 31
SDSU researchers build AI to predict stock market trends and volatility
Researchers from South Dakota State University just unveiled ALERTA-Net, a deep neural network designed to forecast stock price movements and market volatility. By combining macroeconomic data, search queries, and social posts across 41 blue-chip stocks, the model outperformed standard baselines like DP-LSTM. Why it matters: Multi-source inputs could finally help financial models capture sudden market sentiment shifts instead of relying solely on past charts.

Mon20

ResearchReach 38
AI spots pancreatic cancer on regular CT scans
Researchers just introduced PANDA in Nature Medicine, an AI model that spots pancreatic cancer on routine non-contrast CT scans—a feat long considered impossible for radiologists. In real-world tests, it achieved 99.9% specificity and beat human specialists in sensitivity by 34.1%. Why it matters: It turns everyday chest and abdominal CTs into an early-warning net for the deadliest cancer.

Thu16

ResearchReach 57
Meta unveils Emu Video and Emu Edit for precise image editing and video generation
Meta just announced Emu Edit and Emu Video, two generative AI research breakthroughs built on its Emu foundation model. Emu Edit lets users modify images via text instructions while touching only the relevant pixels, while Emu Video uses a streamlined diffusion setup to generate four-second clips. The tech is currently research-only.

Tue14

ResearchReach 38
DeepMind's GraphCast predicts 10-day weather in under a minute
Google DeepMind just introduced GraphCast, an AI weather model published in Science that delivers 10-day global forecasts in under one minute on a single TPU. It beats the gold-standard HRES simulation system on over 90% of test variables and identifies severe storms days earlier. The code is open source and already undergoing live trials at ECMWF.

Tue31

ResearchReach 36
Google DeepMind previews next-gen AlphaFold that folds DNA and drugs
Google DeepMind and Isomorphic Labs just previewed the next generation of AlphaFold, expanding beyond proteins to model DNA, RNA, and drug ligands with atomic accuracy. It outperforms standard docking methods without needing a reference structure, clearing a path for faster drug design.

Mon30

5:50 AMResearchReach 33
AI can diagnose type 2 diabetes from a 10-second voice clip
Researchers just trained an AI model that detects type 2 diabetes by listening to someone speak for just 6 to 10 seconds. Developed by Canadian scientists at Klick Labs, the tool picks up subtle vocal pitch and intensity variations invisible to the human ear, reaching up to 89% accuracy in women and 86% in men. Why it matters: If validated further, voice-based screening could replace painful, costly in-person blood tests with a quick smartphone recording.

Mon23

8:30 PMResearchReach 31
AI uses Google Street View to inspect buildings in hours, not months
University of Florida researchers just developed a deep learning system powered by NVIDIA DGX A100 systems that evaluates urban building safety from Google Street View images. The model slashes post-disaster inspection times from months to a few hours while hitting over 85% accuracy under FEMA standards. It's already being tested across several cities in Mexico.

Sat21

2:11 AMResearchReach 31
Researchers unveil 3D-GPT to turn simple text into 3D worlds
Researchers from ANU, Oxford, and BAAI just developed 3D-GPT, an AI system that translates text prompts into 3D environments. By deploying specialized agents to conceptualize scenes and write code directly for Blender, it automates procedural 3D modeling. It isn't photorealistic yet, but it signals a major shift toward text-driven asset creation.

Wed18

ResearchReach 57
Meta's new AI can decode images directly from brain activity in real time
Meta researchers just showed off an AI system that decodes human brain activity into real-time images using non-invasive magnetoencephalography (MEG). By aligning brain signals with its DINOv2 vision model, the setup reconstructs what a person sees on the fly. It accurately captures object types, though object placement and fine details still struggle.

Wed11

3:00 PMResearchReach 36
Scientists build AI that predicts viral mutations before they hit
Researchers from Harvard and Oxford just introduced EVEscape, an AI tool designed to forecast how viruses will mutate before new variants actually emerge. By combining evolutionary sequences with structural biology, the system spots dangerous escape mutations in SARS-CoV-2 and flu, helping scientists design future-proof vaccines.

Fri6

ResearchReach 31
Meta and Inria's Stable Signature bakes invisible watermarks into AI models
Meta FAIR and Inria just open-sourced Stable Signature, a new watermarking method for open-source AI image generators. Instead of slapping watermarks on exported images, it roots an invisible signature directly inside the model's decoder, making it survive cropping, compression, and ControlNet fine-tuning. Why it matters: it virtually eliminates false positives while tracing deepfakes straight back to the model that made them.

Tue3

10:00 AMResearchReach 33
Researchers put AI watermarks to the test—and broke every single one
University of Maryland researchers just revealed that every existing AI image watermarking method can be cracked. Attackers can easily wipe invisible watermarks clean or even inject them into human-made photos to trigger false positives. While tech giants are betting big on watermarking to fight deepfakes, it's far from a silver bullet.

Wed27

1:00 PMResearchReach 31
AI brings 3-minute cancer biopsy results straight to the operating room
Invenio Imaging just put deep learning right inside the operating room, using NVIDIA edge AI to analyze fresh biopsies in three minutes instead of waiting weeks for a pathology lab. The system achieved 93% accuracy in 90 seconds for tumor biomarkers, letting surgeons make immediate calls during surgery.

Mon25

ResearchReach 31
John Carmack and Rich Sutton team up to build AGI by 2030
Doom creator John Carmack's startup Keen Technologies just partnered with reinforcement learning pioneer Rich Sutton to accelerate artificial general intelligence. Both researchers aim to bypass current mainstream trends by focusing on agency and temporal learning, targeting a genuine AGI prototype by 2030.