20241016 – Alibaba | Tsinghua University | HKU | SenseTime | more
1. Alibaba International releases its first commercial translation large model, the e-commerce tool for global expansion has truly arrived!
Alibaba International has unveiled its new AI translation model, Marco-MT, which promises superior performance compared to existing tools like Google Translate and GPT-4. Specifically designed for e-commerce, Marco-MT excels in providing accurate translations that capture cultural nuances and contextual meanings, significantly enhancing cross-border communication. With its competitive pricing and seamless integration into Alibaba’s platforms, Marco-MT aims to revolutionize the translation landscape in the global e-commerce sector.
Original Chinese article: https://mp.weixin.qq.com/s/sgb6b5dNDXpPG-EruoaeKA
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2Fsgb6b5dNDXpPG-EruoaeKA
2. AI Protein Wins Nobel Prize, Tsinghua’s Nie Zaiqing: Large Models Decode Biological Language.
The recent Nobel Prize in Chemistry awarded to three scientists highlights the pivotal role of AI and computational methods in protein design and structure prediction, signaling vast potential for AI in biomedicine. Professor Nie Zaiqing from Tsinghua University discusses his innovative work on utilizing advanced natural language processing technologies to build a foundational model that bridges biological and natural languages, aiming to enhance drug development efficiency and success rates through human-AI collaboration. The ongoing development of the ChatDD interactive research assistant is set to revolutionize how experts engage with AI, streamlining processes from project initiation to clinical trials.
Original Chinese article: https://mp.weixin.qq.com/s/8dmUMfyxNEdFjk56oIBlFw
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2F8dmUMfyxNEdFjk56oIBlFw
3. LightRAG: The HKU Huang Chao team developed a simple and efficient RAG system, significantly reducing the retrieval enhancement costs of large models.
The University of Hong Kong’s Huang Chao team has introduced LightRAG, an innovative open-source framework that integrates graph structures with a dual-layer retrieval mechanism, significantly reducing retrieval costs while enhancing accuracy and efficiency in information retrieval. By effectively capturing complex relationships between entities and adapting quickly to new data, LightRAG excels in addressing both specific and abstract queries, ensuring users receive contextually relevant responses. This advancement marks a substantial improvement in retrieval-augmented generation (RAG) systems, showcasing its potential for dynamic environments.
Original Chinese article: https://www.jiqizhixin.com/articles/2024-10-14-2
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.jiqizhixin.com%2Farticles%2F2024-10-14-2
4. RMB 69,800! LimX Dynamics’ Bipedal Robot TRON 1 Now Available for Pre-sale.
LimX Dynamics has launched its first multi-form bipedal robot, TRON 1, priced at RMB 69,800. Despite having only a lower body, TRON 1 boasts exceptional mobility and stability, offering three interchangeable modes: dual-point, bipedal, and dual-wheeled. Designed for researchers, it provides an open SDK, Python-based development, and Sim2Real one-click deployment to ease reinforcement learning research. Notably, LimX Dynamics is the first humanoid robot company backed by Alibaba, marking a significant milestone in the AI and robotics industry.
Original Chinese article: https://mp.weixin.qq.com/s/YXtN-ZEsOaiTkhAziEikZg
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2FYXtN-ZEsOaiTkhAziEikZg
5. SenseTime Wins Gold in the SuperCLUE-V Multimodal Large Model Benchmark for October.
SenseTime has once again achieved a remarkable milestone with its multi-modal model, SenseChat-Vision 5.5, which has been ranked at the forefront of China’s AI landscape according to the October SuperCLUE-V benchmark. Scoring 73.56, it not only excelled in multiple tasks but also outperformed all evaluated models, including the latest GPT-4o, in mathematical reasoning capabilities. This recognition underscores SenseTime’s continuous commitment to advancing AI technology and enhancing cognitive abilities within its models.
Original Chinese article: https://www.sensetime.com/cn/news-detail/51169035?categoryId=72
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.sensetime.com%2Fcn%2Fnews-detail%2F51169035%3FcategoryId%3D72
6. Doubao’s Large Model Recreates Blockbuster: Douyin’s “AI Picture Book Special Effects” Exceeds 1.1 Billion Plays.
The “AI Healing Picture Book” effect on Douyin (Chinese version of TikTok) has surpassed one million submissions during the Golden Week, captivating users with its charming, high-fidelity design and innovative 3D page-turning format. Utilizing ByteDance’s Doubao large model, the effect offers impressive customization through advanced algorithms, allowing users to become the protagonists of their own illustrated stories. With over 1.1 billion views on relevant topics, this effect highlights the growing influence of AI in creative media.
Original Chinese article: https://www.qbitai.com/2024/10/199676.html
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.qbitai.com%2F2024%2F10%2F199676.html
7. China’s Large Model Arena Surpasses GPT-4 for the First Time! [01.AI](http://01.ai/) and Zhipu AI’s GLM Both Enter the Top 10.
The latest release of the Yi-Lightning model by the domestic AI firm [01.AI](http://01.ai/) has made waves by surpassing GPT-4o, positioning itself sixth overall in the global Chatbot Arena rankings. This achievement highlights significant advancements in model performance, particularly in mathematics and coding, while offering competitive pricing at just 0.99 RMB per million tokens, thus underlining China’s growing prowess in AI technology.
Original Chinese article: https://mp.weixin.qq.com/s/-z62w1IyTwfrpXvNujFp0Q
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2F-z62w1IyTwfrpXvNujFp0Q
8. Empowered by the AI agent engine: MediaTek’s Dimensity 9400 brings the “complete” AI smartphone to market ahead of schedule.
The recently unveiled MediaTek’s Dimensity 9400 chip is poised to revolutionize mobile technology by integrating advanced AI capabilities directly into smartphones, enabling local AI training and high-quality video generation. This marks a significant leap in mobile processing power, allowing for a seamless user experience that merges various applications and enhances human-machine interaction. With major brands like Vivo set to launch products featuring this cutting-edge chip, the future of AI-powered smartphones looks promising.
Original Chinese article: https://mp.weixin.qq.com/s/FuGX3XcI0_eMYhyWCM2TjQ
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2FFuGX3XcI0_eMYhyWCM2TjQ
9. Baidu’s Short Drama Launches Agent Dialogue Feature in Beta.
Baidu’s Short Drama has launched the agent dialogue feature on October 14, allows users to interact in real-time with characters from short dramas, enhancing the viewing experience by transforming passive consumption into active engagement. Available within the Baidu app, this feature offers users the opportunity to converse with multiple characters simultaneously, creating a dynamic and immersive environment. As the company continues to prioritize the development of short dramas and AI applications, this innovation could redefine content interaction and user engagement in the digital landscape.
Original Chinese article: https://news.qq.com/rain/a/20241015A04UFO00
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fnews.qq.com%2Frain%2Fa%2F20241015A04UFO00
10. WeRide has released the new generation Robotaxi GXR: equipped with L4 level public road operation capability.
WeRide has officially launched its next-generation Robotaxi, the GXR, leveraging over 1,800 days of operational experience in the field. This L4-capable vehicle is designed to navigate complex urban environments seamlessly, equipped with a state-of-the-art HPC2.0 platform providing over 1,300 TOPS of AI computing power and a comprehensive sensor suite for enhanced perception. With innovative features like a hidden B-pillar design and optimized space, the GXR signifies a new era in autonomous transportation, offering users a chance to experience rides through the WeRide Go App.
Original Chinese article: https://www.aibase.com/zh/news/12458
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.aibase.com%2Fzh%2Fnews%2F12458
That’s all for today’s China AI Native Industry Insights. Join us at AI Native Foundation Membership Dashboard for the latest insights on AI Native, or follow our linkedin account at AI Native Foundation and our twitter account at AINativeF.