SenseTime recently launched its new app, 'Magic Photo Capture', providing users with a creative and fun AI imaging experience. With the arrival of the Lunar New Year, this app aims to help users create personalized New Year photos using advanced artificial intelligence technology, enhancing the festive atmosphere. 'Magic Photo Capture' is not just an ordinary photography app; it includes a variety of templates in its mini-program, allowing users to choose from graduation photos, creative group photos, classic art style portraits, travel photos, new national style, and more.

ByteDance officially launches its latest Doubao large model 1.5 Pro (Doubao-1.5-pro), which demonstrates outstanding comprehensive capabilities in various fields, successfully surpassing the well-known GPT-4o and Claude3.5Sonnet in the industry. The release of this model marks an important step forward for ByteDance in the field of artificial intelligence. Doubao 1.5 Pro adopts a novel sparse MoE (Mixture of Experts) architecture, utilizing a smaller set of activation parameters for pre-training. This design's innovation...

On January 22, 2025, Volcano Engine, a subsidiary of ByteDance, officially announced the release of Doubao Large Model 1.5, which is now fully available on the Volcano Ark platform. The newly released Doubao Large Model 1.5 has achieved significant performance improvements across multiple domains, reaching a globally leading level of comprehensive capabilities, marking another important breakthrough for ByteDance in the field of artificial intelligence.

Recently, the DeepSeek team in China has launched its latest open-source large model R1, which has gained widespread attention. The performance of the R1 model is exceptionally outstanding, surpassing OpenAI's o1 model in multiple tests, especially excelling in assessments related to mathematics and programming. In the latest AIME2024 test in the United States, R1 scored 79.8, outperforming o1's score of 79.2. In the MATH-500 test, R1 achieved a score of 97.3, again leading over o1.

DeepSeek has officially released and open-sourced its latest large language model R1, which has shown outstanding performance and is considered comparable to OpenAI's official version o1. This initiative not only marks another significant breakthrough in domestic AI technology but also provides new options for AI developers worldwide. DeepSeek R1 has extensively applied reinforcement learning techniques in the post-training phase, significantly enhancing the model's reasoning capabilities even with minimal labeled data, in areas such as mathematics, coding, and natural language reasoning.

Recently, Zhejiang University and Alibaba DAMO Academy jointly released a remarkable study aimed at creating high-quality multimodal textbooks from educational videos. This innovative research not only provides new ideas for training large-scale language models (VLMs) but may also change the way educational resources are utilized. With the rapid development of artificial intelligence technology, the pre-training corpus of VLMs mainly relies on visual-text pairs and visually intertwined data. However, much of this current data comes from the web, where the correlation between text and images is weak, and the knowledge density is relatively low.

Recently, Sinopec officially launched its first 'AI Digital Employee'. This virtual assistant will be piloted at more than 40 gas stations nationwide, including the Xinyang Jianneng Station in Nanning, Guangxi. As the first AI gasoline digital employee in China's oil and petrochemical industry, this innovative initiative aims to enhance the self-service fueling experience for customers. During the fueling process, drivers can interact with the digital employee through the gas pump's large screen. For example, a driver might ask, 'Hello! I would like to fill up with 95 octane gasoline, what is the price?' The digital employee can quickly respond with the price, and

Recently, Doubao Company announced the launch of its new real-time voice large model, claiming to achieve a "cliff-like lead" in Chinese dialogue, marking a significant enhancement in AI conversational capabilities. This model is fully available in the Doubao App (version 7.2.0 New Year Edition), providing users with a richer and more authentic voice communication experience. According to reports, Doubao's real-time voice large model has achieved a deep integration of speech understanding and generation, forming an end-to-end voice dialogue system. This technological breakthrough allows the model to excel in vocal expressiveness, control, and emotion.

On January 20th, SenseTime announced that its 'SenseNova-5o Interactive Version' is officially providing real-time audio and video conversation services, available to users for free for a limited time. The release of this interactive version marks another significant advancement for SenseTime in the field of artificial intelligence, aiming to enable more natural and smoother communication between AI and humans through powerful real-time interaction, visual recognition, memory thinking, and complex reasoning capabilities.

Recently, iFLYTEK Co., Ltd. has made significant strides in the field of intellectual property by applying for the registration of multiple trademarks related to "Fireworks Large Model", "Torch Large Model", and "Starlight Large Model". According to Tianyancha information, the international classifications of these trademarks cover various fields, including scientific instruments, website services, and education and entertainment. Currently, the status of these trademarks is awaiting substantive examination. Established in December 1999, iFLYTEK is a leading company in the field of artificial intelligence and speech recognition in China, with Liu Qingfeng as the legal representative and a registered capital of approximately 23.1 million.

The AI image detection system developed by Tencent's Hongyuan Security Team's Zhuque Laboratory can determine if an image is AI-generated within a few seconds by uploading the image and waiting for verification. The system primarily distinguishes between real images and AI-generated images by capturing differences, such as AI-generated images sometimes defying common sense logic, showing unrealistic content like dogs flying with wings or cats smoking cigars.

In today's rapid development of Large Language Models (LLM), the training and inference costs of models have increasingly become a focus of research and application. Recently, the Tencent Hunyuan team released an important study that delves into the 'Scaling Laws' of low-bit floating point quantization training, which refers to the principles governing the scale of floating point quantization training. The core of this research lies in exploring how to significantly reduce computational and storage costs without sacrificing performance by lowering the precision of the model.

According to reports, the U.S. Department of Commerce has added Zhipu AI to its entity list for export controls, accusing it of promoting China's military modernization and infringing on U.S. national security and diplomatic interests. ** This decision makes Zhipu AI the first Chinese large model company to be added to the U.S. entity list, marking a new round of suppression by the U.S. on China's artificial intelligence industry. Zhipu AI subsequently issued a statement emphasizing that the company is based on technology transfer from Tsinghua University and provided a detailed review of its technological development history. Zhipu AI started developing the GLM pre-training architecture in 2020 and launched it in 2022.

Recently, Alibaba Damo Academy launched a multi-modal large language model named Valley 2, designed for e-commerce scenarios. It aims to enhance performance across various fields and expand the application boundaries of e-commerce and short video scenarios through a scalable vision-language architecture. Valley 2 utilizes Qwen 2.5 as its LLM backbone, paired with the SigLIP-384 visual encoder, incorporating MLP layers and convolution for efficient feature transformation.

New developments in the field of artificial intelligence! Former ByteDance large model solutions architect Pan Yunan has officially joined the startup Lobo Intelligence as co-founder and CTO. This news has attracted widespread attention in the industry, signaling that Lobo Intelligence will accelerate its efforts in the integration of large models and embodied intelligent hardware applications. Founded in January 2024, Lobo Intelligence focuses on innovative applications of large models and embodied intelligent hardware. It is reported that the company has completed angel financing amounting to tens of millions of RMB, which will support subsequent product research and development.

Recently, the AI company SBot announced the completion of a new round of financing, amounting to 500 million RMB. The participants in this round of financing include well-known industrial funds, state-owned platform investments, and various private equity funds. The infusion of this capital will further promote SBot's rapid development in smart terminals and industry applications. Image source note: The image generated by AI is authorized by the service provider Midjourney. It is understood that the success of this financing is mainly due to the company's scalable commercial capabilities in end-side application scenarios, as well as its advancements in large model human-computer dialogue technology.

In recent years, the programming capabilities of Artificial Intelligence (AI) have been continuously evolving, but they have not yet reached perfection. Recently, Max Woolf, a senior data scientist at BuzzFeed, discovered through experiments that if continuous prompts for 'writing better code' are provided to large language models (LLMs), AI can indeed generate higher quality code. This finding has garnered widespread attention, with renowned AI scientists emphasizing the importance of iteration and prompt design. In Woolf's experiment,