Welcome to the AI Daily section! Here, you'll find your daily guide to exploring the world of artificial intelligence. Every day, we bring you the hottest topics in the AI field, focusing on developers to help you understand technological trends and innovative AI product applications.
Fresh AI Products Click to Learn More: https://top.aibase.com/
1. Apple WWDC Unleashes a Deep Dive: GPT-4o Powers Siri, Bringing Generative AI to the Entire Family
Apple announced at its 2024 Worldwide Developers Conference (WWDC) that all its products are entering the era of generative AI, introducing the new personalized smart system, Apple Intelligence. The core update combines generative AI models with user profiles for intelligent services, deeply integrated into iOS18, iPadOS18, and macOS Sequoia. Siri undergoes a transformation, boasting enhanced language understanding capabilities and cross-application operation execution. The system integrates ChatGPT to provide image and document understanding functions, along with new writing tools and Image Playground features.
AiBase Highlights:
🍎 Apple Intelligence integrates generative AI models and user profiles to provide practical smart services, deeply integrated into iOS18, iPadOS18, and macOS Sequoia.
🤖 Siri undergoes a transformation, with richer language understanding capabilities, supporting cross-application operations, and allowing users to interact with Siri via typing.
📸 The system integrates ChatGPT to provide image and document understanding functions, along with new writing tools and Image Playground, enabling users to create animations, illustrations, or sketch-style images.
2. Apple Collaborates with Google's Gemini Model
Apple announces collaboration with Google's Gemini model, opening up third-party model integration to offer users more choices. Siri will integrate with ChatGPT, allowing users to engage in conversations without leaving Siri while maintaining privacy controls. Apple updates its development toolkit,首次接入 OpenAI's ChatGPT, and releases a series of new features and updates.
AiBase Highlights:
🍎 Apple collaborates with Google's Gemini model, opening up third-party model integration to expand the AI ecosystem.
🤖 Siri integrates with ChatGPT, allowing users to converse within Siri while maintaining privacy controls.
🚀 Apple updates its development toolkit,首次接入 OpenAI's ChatGPT, and releases new features for iOS18 and VisionOS2.
3. iOS18 Photo App Overhaul: New AI Removal Feature and Intelligent Screening to Narrow Search Scope
Apple's latest iOS18 system features a comprehensive overhaul of the messaging function. Users can now add underlines and strikethroughs to message content and apply a series of dynamic text effects, making each message unique.
AiBase Highlights:
⭐️ Apple reaches a collaboration agreement with OpenAI, introducing ChatGPT functionality to iOS18.
🤖 Insights generated by GPT-4 are informative for future stock performance.
💬 iPadOS18 supports various customization features on iOS18 and introduces custom function bars within applications.
Check out the iOS18 upgrade compatibility list here: https://www.chinaz.com/2024/0611/1622488.shtml
4. Tencent Launches New Image-to-Video Model Follow-Your-Pose-v2
This article introduces Tencent's new image-to-video model "Follow-Your-Pose-v2," developed in collaboration with Sun Yat-sen University and the Hong Kong University of Science and Technology. The model achieves multi-person video action generation, strong generalization capabilities, and correct handling of character occlusions, marking significant progress in the field of video generation with broad application prospects.
AiBase Highlights:
🌟 Supports multi-person video action generation, with less inference time.
🔥 Strong generalization capabilities, able to generate high-quality videos regardless of age, clothing, race, background clutter, or action complexity.
💡 Correctly handles character occlusions, generating scenes with correct foreground-background relationships.
Project Page: https://top.aibase.com/tool/follow-your-pose
Paper Address: https://arxiv.org/pdf/2406.03035
5. MotionFollower: Recreating Human Motion Without Altering Video Background
MotionFollower is an innovative technology that can copy motion from one video onto a character in another video while maintaining the character's appearance unchanged. This technology has a wide range of applications in film production, advertising creation, game development, and more.
AiBase Highlights:
⚙️ MotionFollower is an innovative technology that copies motion from one video onto another character while maintaining the appearance unchanged.
🌐 Widely applicable in film, advertising, gaming, and more.
🎥 MotionFollower handles videos with extensive camera movements, achieving high-quality motion transfer.
Details Link: https://top.aibase.com/tool/motionfollower
6. Adobe Revises Service Terms, Clarifying No Use of Customer Works for AI Training
Adobe announces revised service terms, clarifying that customer works will not be used for AI training, aiming to regain user trust. This change comes after strong user protests a week earlier.
AiBase Highlights:
🛡️ Adobe revises service terms, clarifying that customer works will not be used for AI training.
💬 Adobe's president acknowledges the need for earlier clarification of service terms and pledges greater transparency.
🖼️ Creators' concerns about Adobe remain, as the company strives to regain trust.
7. OpenAI Upgrades ChatGPT Voice Functionality, Allowing It to Speak in Different Character Voices
OpenAI's latest update to ChatGPT's voice functionality allows users to interact with the chatbot using various AI-generated voices and styles. The new feature enhances interactivity and accessibility by enabling users to instruct the AI chatbot to respond in any voice in real-time.
AiBase Highlights:
🔊 ChatGPT now offers four preset voices, with real-time optimization of voice styles.
🗣️ Users can request AI to dub characters in stories, generating unique voices, such as a lion's roar.
🔜 OpenAI will roll out new voice features in the coming weeks to all ChatGPT users, with premium subscribers getting priority access.
8. Surpassing Instant3D! Shanghai Jiao Tong University Introduces New Framework Bootstrap3D, Significantly Enhancing 3D Generation Capabilities
A research team from Shanghai Jiao Tong University and the Chinese University of Hong Kong has introduced a new framework called Bootstrap3D, which combines fine-tuned 3D-aware multimodal large models to automatically generate high-quality multi-view image data, significantly enhancing the capabilities of 3D generation models. The framework's synthetic dataset is fully open-source for researchers and developers to use for free. Key features of the framework include a data construction pipeline, text prompt generation, image generation, multi-view synthesis, quality screening, and description rewriting. The research team also proposes a training timestep reordering (TTR) strategy to optimize different stages of the denoising process, addressing issues in multi-view diffusion model training. Experimental results show that multi-view diffusion models using the TTR strategy excel in image-text alignment, image quality, and view consistency, effectively improving multi-view generation results.
AiBase Highlights:
🔑 Data Construction Pipeline: Automatically generates multi-view image data and detailed description text, a core innovation of the framework.
🔑 Text Prompt Generation: Uses large language models to generate creative and diverse text prompts for image generation.
🔑 Multi-View Synthesis: Expands single-view images into multi-view images, ensuring consistency across different views.
Details Link: https://top.aibase.com/tool/bootstrap3d
9. Google Introduces AGREE Framework to Enhance Accuracy of Large Language Models' Generated Content
Google Research introduces the AGREE framework, aimed at enhancing the accuracy of content and references generated by large language models. The framework improves answer accuracy by retrieving relevant paragraphs and provides users with a way to verify the authenticity of information. Core technologies include fine-tuning during training and adaptive adjustments during testing. Experimental results show that AGREE significantly improves the accuracy and reference quality of content responses.
AiBase Highlights:
🔍 The AGREE framework aims to enhance the accuracy of content and references generated by large language models.
🎯 Core technologies include fine-tuning during training and adaptive adjustments during testing.
💡 Experimental results show that AGREE significantly improves the accuracy and reference quality of content responses.
Details Link: https://arxiv.org/pdf/2311.09533
10. Fenbi to Launch Self-Developed AI Smart Teacher in August
Fenbi Group will launch its self-developed AI Smart Teacher in August 2024, becoming one of the AI learning tools offered on its online platform, initially applied to national or provincial recruitment and certification exam system classes.
AIBase Highlights:
⭐️ Fenbi Group will launch its self-developed AI Smart Teacher in August 2024.
⭐️ The AI Smart Teacher will be one of the AI learning tools offered on Fenbi's online platform.
⭐️ Initially applied to national or provincial recruitment and certification exam system classes.
