Welcome to the AI Daily section! Here, you'll find your daily guide to exploring the world of artificial intelligence. Each day, we bring you the hottest topics in the AI field, focusing on developers to help you understand technology trends and discover innovative AI product applications.
Fresh AI Products Click to Learn More: https://top.aibase.com/
1. Kuaishou Launches "Keling AI" Standalone App to Expand Mobile AI Creation Layout
Kuaishou recently launched the "Keling AI" standalone app, dedicated to expanding the AI creation field on mobile devices. The app is based on Kuaishou's self-developed Keling and Ketu models, offering video and image generation and editing capabilities. Although it lacks some features compared to the web version, it has attracted 3.6 million users and created over 37 million videos and one billion images. It competes with Douyin's "Dream AI" app.
AiBase Highlights:
🚀 Kuaishou launches "Keling AI" standalone app, expanding mobile AI creation layout.
💡 "Keling AI" app is based on self-developed models, offering video and image generation and editing capabilities.
💪 Attracted 3.6 million users, created over 37 million videos and one billion images. Competes with Douyin's "Dream AI".
2. Zhipu AI Open-Sources Latest Version of Video Model CogVideoX v1.5
Zhipu Technology team has released the latest video generation model CogVideoX v1.5 and open-sourced it. This version has significant improvements in video generation capabilities, supporting 5/10-second videos, 768P resolution, and 16-frame generation capabilities. The Xinqingying platform has launched, combining with the CogSound audio model, offering ultra-high-definition 4K video generation. Data processing and algorithmic innovations ensure video quality and coherence.
AiBase Highlights:
🚀 CogVideoX v1.5 open-sourced, supporting 5/10-second videos, 768P resolution, and 16-frame generation capabilities.
🎬 Xinqingying platform launched, combining with CogSound audio model, offering ultra-high-definition 4K video generation.
🔬 Data processing and algorithmic innovations ensure the quality and coherence of generated videos.
Details link: https://github.com/thudm/cogvideo
3. Report: Baidu to Launch Smart Glasses with Built-in AI Assistant, Based on Ernie Model
Baidu plans to launch smart glasses with a built-in XiaoDu AI assistant, which will be showcased at the 2024 Baidu World Conference. These glasses feature innovative functions, providing a new intelligent experience.
AiBase Highlights:
🌟 The glasses will be equipped with a camera, supporting photo and video shooting functions.
📷 The glasses support voice interaction based on the Ernie model, enhancing user experience.
🗣️ Baidu's smart glasses will provide more personalized and China-user-oriented solutions.
4. Big Spenders Rejoice! Krea AI Launches Lora Training Feature
Krea AI has launched the Lora training feature, offering users personalized training services for characters, styles, products, etc. Users can upload multiple images to start training, with a simple and easy-to-use interface. Training parameter settings are straightforward, and the training results are good, friendly for novice users. Membership costs $10 per month and supports commercial use.
AiBase Highlights:
🔥 Krea AI launches Lora training feature, allowing users to customize training services for characters, styles, products, etc.
💡 Interface is simple, upload images to set basic parameters, training process is easy to use.
💰 Membership price is $10 per month, can generate 720 FLux images, 36,000 real-time images, supports commercial use.
Details link: https://www.krea.ai/train
5. Keling Launches Custom Model Feature to Overcome AI Video Generation Character Consistency Issues
Keling's latest custom model feature brings an important breakthrough to the AI video generation field, solving the technical problem of character consistency. This feature allows creators to upload multiple video clips for training, supports using AI videos generated by Keling itself, providing greater flexibility. After training, the system can generate video content that maintains character consistency, showcasing a colorful range of creative possibilities.
AiBase Highlights:
🔑 Creators can upload multiple video clips for model training, supports using AI videos generated by Keling itself, providing greater flexibility.
🚀 After training, the system can generate video content that maintains character consistency, showcasing 360-degree rotation and stable appearance features in complex scenes.
💡 KLING1.5 version supports 1080p HD video output, offers multiple format choices, meets content needs for different platforms. Action brush tools enhance video creation precision and flexibility.
6. Compact and Powerful! Pocket-Sized Visual AI Model moondream2
Moondream's moondream2 is an outstanding compact visual language model with only 1.6 billion parameters, capable of running on small devices like smartphones. The model has powerful text and image processing capabilities and performs excellently in various benchmark tests. The Moondream team continues to update the model's performance, successfully raising $4.5 million in funding, with the CEO having extensive experience.
AiBase Highlights:
🌟 Moondream launches moondream2, with only 1.6 billion parameters, capable of running on smartphones and other small devices.
💰 Moondream successfully raises $4.5 million in funding, the CEO previously worked at Amazon, the team continues to update and improve the model's performance.
📈 moondream2 has powerful text and image processing capabilities, can answer questions, perform OCR, object counting and classification, performs excellently in benchmark tests.
Details link: https://huggingface.co/vikhyatk/moondream2
7. Xiaomi Announces Big Model XiaoAi Upgrade: Multi-Device Support for "Music Q&A" Feature
Xiaomi officially announces the upgrade of the big model XiaoAi with the addition of the "Music Q&A" feature, allowing users to experience it on multiple devices. This feature is based on the enhancement of artificial intelligence capabilities, helping users query song information and answer professional music knowledge. XiaoAi went through a major version update in July this year, comprehensively improving natural Q&A and smart assistant functions, enhancing user experience.
AiBase Highlights:
🎵 Xiaomi's XiaoAi adds "Music Q&A" feature, supports song information query and professional music knowledge answers.
📱 Feature is applicable to multiple devices, users need to upgrade XiaoAi to the latest version.
🚗 In past updates, XiaoAi has comprehensively improved natural Q&A and multiple smart assistant functions, enhancing user experience.
8. Mistral AI Launches Multilingual Content Moderation API, Challenging OpenAI
Mistral AI has launched a new content moderation API aimed at competing with industry leaders, addressing AI safety and content filtering issues. The API is based on the Ministral8B model, capable of detecting various harmful content, supports multiple languages, and has been deployed on the Le Chat platform. Mistral has partnered with multiple enterprises, showcasing its mature technology and European perspective.
AiBase Highlights:
🌍 Multilingual support: API covers 11 languages, detects pornographic content, hate speech, and other harmful content.
🔗 Partnerships: Applied on Le Chat platform, established partnerships with Microsoft Azure, Qualcomm, etc.
🔒 Technical advantages: Based on dialogue context, enhances detection capabilities for potentially harmful content.
Details link: https://docs.mistral.ai/capabilities/guardrailing/
9. AI "Electronic Tongue" Here! Can Taste Food Flavors and Easily Detect Food Freshness
Scientists have successfully developed a new device called the "electronic tongue," using artificial intelligence technology to accurately distinguish different coffee blends and detect the freshness of food and beverages, providing new solutions for food testing and quality control. The electronic tongue mimics human taste technology, improving accuracy and efficiency, accelerating flavor development and ensuring product quality. It has the function of monitoring food freshness, protecting consumer health, and quality control for businesses. The future application prospects are broad, potentially extending to multiple fields such as pharmaceuticals and environmental monitoring.
AiBase Highlights:
🌟 Electronic tongue uses artificial intelligence technology to accurately distinguish different coffee blends.
🍽️ Detects the freshness of food and beverages, preventing spoilage.
🔬 The future electronic tongue is expected to be applied in multiple fields such as pharmaceuticals and environmental monitoring.
Details link: https://www.nature.com/articles/s41586-024-08003-w
10. Google's Black Technology: ReCapture Technology Analysis
ReCapture technology is an innovative technology launched by Google Research, allowing users to experience the original video content from a new perspective by regenerating videos with custom camera trajectories. The principle behind this technology is not complex, using multi-view diffusion models and mask video fine-tuning technology to transform rough videos into clear, coherent, and dynamic videos. Google researchers say that ReCapture does not require a lot of training data and is suitable for various videos and perspective transformations, allowing ordinary users to easily create professional-grade "multi-camera" videos.
AiBase Highlights:
🔍 ReCapture technology allows users to re-experience videos, generating new versions with custom camera trajectories.
🛠️ Uses multi-view diffusion models and mask video fine-tuning technology to transform rough videos into clear, coherent, and dynamic videos.
🎥 ReCapture does not require a lot of training data, suitable for various videos and perspective transformations, allowing ordinary users to create professional-grade "multi-camera" videos.
Details link: https://generative-video-camera-controls.github.io/
11. OpenAI Wins! New York Court Dismisses Raw Story Copyright Lawsuit, Is AI Data Scraping Legalization a Trend?
Recently, the Southern District Court of New York dismissed the copyright infringement lawsuit filed by Raw Story Media, Inc. and AlterNet Media, Inc. against OpenAI, finding that the case lacked standing. This ruling marks an important victory for OpenAI in the legal disputes between AI and copyright, providing a reference for future similar cases. The court's decision to dismiss the lawsuit emphasizes that the comprehensive nature of generative AI content makes verbatim copying less likely, providing legal reference and enlightenment for future handling of AI and copyright disputes.
AiBase Highlights: