Generating high-quality, temporally continuous videos requires a significant amount of computational resources, especially for longer time spans. The latest Diffusion Transformer model (DiTs) has made significant progress in video generation, but its reliance on larger models and more complex attention mechanisms has resulted in slower inference speeds, exacerbating the challenge. To address this issue, researchers at Meta AI have proposed a training-free method called AdaCache to accelerate video DiTs.

In recent years, non-destructive imaging technology has made rapid advancements in the field of painting research and preservation. Among these, macro X-ray fluorescence (MA-XRF) analysis stands out, helping experts identify pigments and analyze painting techniques, providing valuable insights into the artist's creative process. However, MA-XRF technology generates large and complex datasets that challenge traditional data analysis methods. Recently, Italian researchers applied deep learning algorithms to spectral analysis of MA-XRF datasets, developing a brand new analysis approach.

Tencent Cloud and Xiaomi have collaborated to launch the new Smart Lock M30 Pro. This product utilizes Tencent's 'palm scanning technology' that leverages AI recognition capabilities of palm prints and palm veins, providing users with a novel and convenient way to unlock securely. This technology is based on Tencent's proprietary 'palm print + palm vein' recognition algorithm, which captures images of the veins beneath the palm skin using near-infrared light for high-precision identity verification. Since every individual's palm vein pattern is unique, this technology not only enhances the convenience of unlocking but also significantly increases security.

According to 36Kr, Xiaomi's automotive autonomous driving team recently completed a new round of organizational restructuring, marking the entry of Xiaomi's autonomous driving business into a new acceleration phase. The original two secondary departments, 'Perception' and 'Control', have been merged into the 'End-to-End Algorithm and Function Department', responsible for production-ready solution development. This change indicates that Xiaomi's autonomous driving will shift from traditional modular development to the technological route of large end-to-end models. The restructuring of Xiaomi's autonomous driving business and the shift in technological direction aim to capture user mindset and maintain high sales.

Recently, the AI assistant Flow officially released version 2.0, introducing the new 'Featured Content' and 'Discover' sections that provide users with more intelligent and personalized high-quality content services. These features demonstrate the practical value of artificial intelligence platforms in both life and work, especially in addressing users' confusion and needs. With the continuous maturity of large model technology, Flow offers users a richer variety of information in graphic, text, and Q&A formats through AI Native Content (AINative). The new version's 'Featured Content'

With the rise of generative AI, the prompt trading market is rapidly expanding. However, current trading platforms, exemplified by PromptBase, are predominantly seller-driven in pricing and lack objective price measurement standards. In response to this challenge, the multimedia and intelligent security team at Fudan University has proposed an innovative prompt trading model aimed at better adapting to a future buyer-driven market. This new trading model consists of two main phases: prompt category selection and pricing strategy formulation. In the first phase, the platform employs a greedy search-based multi-armed bandit algorithm to...

Within the complex neural networks of the human brain, billions of neurons are continuously generating electrical activity, encoding our every thought, action, and sensation. The complexity of this neural symphony has long been a significant challenge in brain-computer interface (BCI) research. However, researchers at the University of Southern California (USC) have recently made a major breakthrough by developing a novel artificial intelligence algorithm called DPAD (Decoupled Prioritized Dynamic Analysis), which promises to fundamentally change the way we interpret brain activity. Image Source Note: Image generated by AI, image licensed by provider M

At the 'All-In Summit 2024', Google co-founder and former Alphabet president Sergey Brin expressed his excitement about the latest advancements in AI and stated that he is working at Google almost every day. Brin confirmed his return to Google's headquarters last year to participate in AI projects, emphasizing the value of the rapid development in the AI field. He noted that algorithm improvements may have already surpassed the increase in computational power requirements, maintaining an optimistic view on AI and believing that future AI model training will not require a significant increase in computational capabilities. Brin's return has implications for Google's future development and AI.

Zhejiang University and the University of Texas developed an AI algorithm that accurately predicts earthquakes in China, forecasting 14 quakes within a 300 km range a week in advance with a 70% accuracy rate, including intensity calculations. This breakthrough offers new hope for earthquake warning and disaster response. While global application of the algorithm remains uncertain, scientists are optimistic about improving prediction accuracy and ....

Melodisco is an open-source AI music player that focuses on personalized music experiences. It features a collection of 300,000 AI songs, with capabilities for AI music generation, smart recommendations, and diverse playlist management. Users can explore charts like the trending, popular, and new release charts, and use random roaming to discover new songs. It supports song favorites, recent play history, and import and share functionalities. In terms of player component design, it offers sequential, random, and single-loop play modes, along with progress and volume control. It integrates the Stripe payment system and supports Vercel and Dock.

The medical large model 'Da Yi' developed by SenseTime Medical has successfully passed the algorithm filing by the National Cyberspace Administration. This filing marks the authoritative recognition of 'Da Yi's' compliance and safety, and highlights a key progress in its technological application. The model is based on the large language model 'Ri Ri Xin · Shang Liang' with a scale of hundreds of billions of parameters.

The 2nd MER24 Challenge focused on the application of AI in emotional recognition within real human-computer interaction scenarios, attracting renowned international scholars. The voice technology team from Soul App excelled in the highly challenging Semi track, winning first place. Their success stemmed from innovations in multimodal data understanding, emotional recognition algorithms, model optimization, and efficient collaboration. To tackle ....

PaddleOCR v2.8.0发布，集前沿技术、优化项目结构、解决历史问题与增强性能于一体，树立OCR领域新标准。版本亮点包括冠军算法SVTRv2与SLANet-LCNetV2，优化算法及更新，修复关键问题，引入符合PEP518规范的教程站点。感谢社区支持，PM成员与贡献者功不可没。开发者可通过访问项目地址【https://github.com/PaddlePaddle/PaddleOCR】获取更多资源与信息。

A new study has developed a new AI tool that makes it easier to detect lies, but only some people agree to use it. Traditional lie detectors appear to measure a person's breathing rate, pulse, blood pressure, and sweating to determine if they are lying, but this 85-year-old technology has long been proven incorrect by scientists. Essentially, the possibility of false positives and the subjectivity of interpreting the results greatly weaken the practicality of lie detectors. Notably, their result

Recently, Kunlun Wanwei, in collaboration with Nanyang Technological University in Singapore, successfully developed an algorithm called Q* that significantly enhances the inference capabilities of existing large models. Q* enables smaller models to achieve inference capabilities comparable to those of models with tens or even hundreds of times more parameters across various datasets, leading to a substantial improvement in model performance while dramatically reducing the demand for computational resources. This brings new possibilities for the widespread application of artificial intelligence and ushers in a new era of efficient intelligence.

While large language models like ChatGPT, Claude, and Gemini are impressive, they all share a significant issue: they often generate hallucinations. This is a serious problem in the field of artificial intelligence, and even Apple has expressed concerns about how future Apple Intelligence will handle hallucinations. Fortunately, a team of researchers has now developed an AI hallucination detector that can determine whether the AI is fabricating content.

The algorithm, named DenseAV, learns the meaning of language by associating audio and video signals. After training DenseAV for an audio-visual matching game, Hamilton and his colleagues observed the pixels the model focused on when it heard sounds. For example, when someone says 'dog', the algorithm immediately looks for dogs in the video stream. This selection of pixels can help people understand what the algorithm believes a particular word means.