Zhipu Releases GLM-Omni Multimodal Model Paper
Zhipu AI and Tsinghua University released a paper detailing the GLM-Omni multimodal model. Experiments show a 54% accuracy in image generation versus 14.57% in text QA, highlighting bottlenecks in unified architectures. Speculation links it to GLM-5.5, though unconfirmed.
Summarized by AI from one outlet
Source
Latest
- Research
Google Deepmind releases satellite forest map
Google Deepmind and the World Resources Institute released ForTy, a global map using satellite data to classify forests into six categories. While it aids in tracking deforestation under EU rules, it identifies only about half of planted forests and tree-crop areas.
- Research
GNC and imec expand collaboration on high-NA EUV dry photoresist
GNC and imec have expanded their collaboration to focus on high-NA EUV dry photoresist technology, supporting semiconductor mass production driven by rising AI compute demands. The partnership aims to address critical material challenges in advanced processes to accelerate commercialization.
- Research
Indian Workers Wear Cameras to Train Robots
A new trend in India uses daily footage from low-wage workers to train technology systems. Workers wear head-mounted cameras to record their movements, assisting tech companies in developing and training future robots and AI systems.
- Research
World model assists NPU chip design
A research team used a world model to guide chip design, creating an ultra-large NPU. The device offers 700 TOPS of compute power with 99% of resources utilized, marking a new advance in AI-driven hardware R&D.
- Research
Experts Warn AI Therapy Avatars May Delay Treatment
More individuals are using generative AI to role-play therapists during their psychological treatment. Experts warn that this practice may confuse the therapeutic process and lead patients to delay seeking professional care, potentially harming their mental health.
All news → · All rights belong to the original creators