Search AI Products and News

Google's NotebookLM has launched a new feature, "Discover Sources," designed to help users easily access relevant information from the web. As a powerful learning tool, NotebookLM allows users to describe a topic of interest, and the system quickly finds and summarizes relevant web content. Users can easily add these sources to their notebooks with a single click for convenient access and future use. This new feature offers a streamlined and user-friendly experience. Users in Noteboo...

On December 30th, Alipay launched its next-generation AI visual search product 'Explore', based on self-developed multimodal large model technology, enabling users to 'explore everything with the eye of AI' and providing faster, more useful, and fun generative search services. Users can let AI recognize plants, pets, and trendy products through the camera when encountering interesting items, get on-the-go travel explanations, check product and medication details, and even enjoy fun interpretations of pet and baby photos, eliminating worries about captions when sharing pictures. The product is now available on Alipay; simply swipe left after clicking 'Scan' on the homepage to find it.

Recently, xAI engineer Hieu Pham posted an 'explosive' message on social media claiming that xAI's large language model Grok3 had successfully proven the Riemann Hypothesis, leading to a halt in the model's training. This news quickly sparked heated discussions in the AI community, with netizens expressing shock and struggling to discern the truth. After all, the Riemann Hypothesis is one of the seven Millennium Prize Problems, hailed as the 'crown jewel' of conjectures, its difficulty is well understood. Hours later, Pham revealed in another post that it was just a joke.

Recently, a new study from the University of Michigan found that an energy-efficient method for training large language models can achieve similar results in the same amount of time while reducing energy consumption by 30%. This method can save enough energy to power 1.1 million American households by 2026. The researchers developed a software tool named Perseus that identifies the critical path, which is a series of subtasks requiring the longest time to complete. Then, Perseus reduces the processor speed on non-critical paths so that all tasks can finish simultaneously.

Is AI actually starting to ‘get smarter’?! The latest research from MIT indicates that the internal structure of large language models (LLMs) astonishingly resembles that of the human brain! This study utilized sparse autoencoder techniques to conduct an in-depth analysis of the activation space of LLMs, revealing three hierarchical structural features that are truly remarkable: first, at the microscopic level, researchers discovered the existence of structures akin to ‘crystals’. The surfaces of these ‘crystals’ are composed of parallelograms or trapezoids, similar to familiar lexical analogies such as ‘man:’

KAN2.0, an upgraded version of the neural network architecture, aims to deepen the integration of AI and classical physics. Researchers can customize the model to incorporate expertise for discovering key concepts in physical systems, such as the Lagrangian. The new framework enhances the model's expressiveness, practicality, and interpretability through three main features: MultKAN, kanpiler, and Tree Converter, addressing compatibility issues between AI and science. KAN2.0 excels in feature recognition, revealing module structures, and discovering symbolic formulas, particularly in fields that are difficult to represent with symbolic equations, such as chemistry.

Researchers from Stanford University and the University of California, Berkeley conducted an experiment to explore the consequences of retraining AI image generation models using images generated by the AI itself. The experiment utilized the Stable Diffusion (SD) model and selected high-quality facial images from the FFHQ dataset for automatic generation. The results showed that after retraining, the AI model produced highly distorted images. This distortion originated not only from the text prompts used for retraining but also persisted even when retraining was done solely with real images.

As the consumption model of digital media continues to evolve, personalized content recommendation has become a key way for users to obtain information. Perplexity Company is keeping up with the trend and is undergoing a major update to its Discover network to meet users' demand for diverse thematic content.The core of this update is the introduction of an expanded list of themes, which will allow users to customize their news sources based on their interests. "This is not just a simple update,"

Silicon Valley unicorn Cobalt Metals has discovered giant copper reserves at the Mingoba project in Zambia using AI technology. Copper is a very important metal with a wide range of applications for humanity. Cobalt Metals was founded with investments from billionaires like Bill Gates and plans to use AI technology to locate various metal minerals. The company claims to have found a world-class copper mine in Zambia using AI technology. Critics have pointed out that the company has not disclosed specific details about the application of AI technology, while stating that the Mingoba copper mine was already discovered. The chief scientist of Cobalt Metals has previously published on the use of AI in mineral exploration.

Google DeepMind has discovered 2.2 million theoretically stable but unrealized crystal structures using AI tools, potentially advancing fields such as renewable energy and advanced computing. The plan is to make the most promising 381,000 structures available to other scientists. The new materials discovered using AI can be utilized in multifunctional layered materials and biomimetic computing applications.

The research team at Daegu University of Science and Technology in South Korea has successfully developed a small sample learning model that can accurately classify using only a small amount of brainwave data, representing a revolutionary advancement. This model overcomes the limitation of traditional deep learning models, which require large volumes of data, and is expected to drive new breakthroughs in brainwave-related research. The research team utilized embedding modules, temporal attention modules, aggregation attention modules, and relational modules to enhance the classification accuracy of the model. The new deep learning model demonstrates up to 76% classification accuracy in cross-subject classification, far surpassing previous methods. This research has promising applications in medical fields.