current location:Home > Technical Articles > Technology peripherals > AI
- Direction:
- All web3.0 Backend Development Web Front-end Database Operation and Maintenance Development Tools PHP Framework Daily Programming WeChat Applet Common Problem Other Tech CMS Tutorial Java System Tutorial Computer Tutorials Hardware Tutorial Mobile Tutorial Software Tutorial Mobile Game Tutorial
- Classify:
-
- To improve the utilization of optical data sets, the Tianda team proposed an AI model to enhance spectral prediction effects
- Editor | Dead Leaf Butterfly Recently, the team of Associate Professor Wu Liang and Academician Yao Jianquan of the Institute of Laser and Optoelectronics of Tianjin University and the team of Professor Xiong Deyi of the Natural Language Processing Laboratory reported a solution that uses a deep learning model with multi-frequency supplementary input to enhance the spectral prediction effect. . This scheme can improve the accuracy of spectral prediction by using multi-frequency input data. In addition, this solution can also reduce noise interference in the spectrum prediction process, thereby improving the prediction effect. This solution can improve the utilization of existing optical data sets and enhance the prediction effect of the spectral response corresponding to the metasurface structure without increasing the training cost. Relevant research results are titled "Enhancedspectrumpredictionusingdeep
- AI 649 2024-06-06 12:09:28
-
- A single 4090 inferable, 200 billion sparse large model 'Tiangong MoE' is open source
- In the wave of large models, training and deploying state-of-the-art dense set LLMs poses huge challenges in terms of computational requirements and associated costs, especially at scales of tens or hundreds of billions of parameters. To address these challenges, sparse models, such as Mixture of Experts (MoE) models, have become increasingly important. These models offer an economically viable alternative by distributing computation to various specialized sub-models, or "experts," with the potential to match or even exceed the performance of dense set models with very low resource requirements. On June 3, another important news came from the open source large model field: Kunlun Wanwei announced the open source 200 billion sparse large model Skywork-MoE, which significantly reduces the inference cost while maintaining strong performance. Based on the previous Kunlun Wanwei open source Skywo
- AI 923 2024-06-05 22:14:46
-
- HuggingFace teaches you how to make a SOTA visual model
- There was OpenAI's GPT-4o before, and Google's series of kings followed. Advanced multi-modal large models hit the market one after another. Other practitioners were shocked and began to think about how to catch up with these super models again. In this paper by HuggingFace and Sorbonne University in France, they summarized the key experiences in building large visual models and pointed out a way for developers. These experiences in the pictures cover many aspects such as model architecture selection, training methods, and training data. The author gives a detailed summary after multiple comparisons. The core points include: If you want to do a good job in large visual models, the choice of architecture is very important. The language model has a greater impact on overall performance than the visual module. Adopting a staged pre-training strategy is more conducive to building model capabilities. The training data should include
- AI 933 2024-06-05 21:39:58
-
- Five schools of machine learning you don't know about
- Machine learning is an important branch of artificial intelligence that gives computers the ability to learn from data and improve their capabilities without being explicitly programmed. Machine learning has a wide range of applications in various fields, from image recognition and natural language processing to recommendation systems and fraud detection, and it is changing the way we live. There are many different methods and theories in the field of machine learning, among which the five most influential methods are called the "Five Schools of Machine Learning". The five major schools are the symbolic school, the connectionist school, the evolutionary school, the Bayesian school and the analogy school. 1. Symbolism, also known as symbolism, emphasizes the use of symbols for logical reasoning and expression of knowledge. This school of thought believes that learning is a process of reverse deduction, through existing
- AI 787 2024-06-05 20:51:22
-
- Bytedance Beanbao large model released, Volcano Engine full-stack AI service helps enterprises intelligently transform
- Tan Dai, President of Volcano Engine, said that companies that want to implement large models well face three key challenges: model effectiveness, inference costs, and implementation difficulty: they must have good basic large models as support to solve complex problems, and they must also have low-cost inference. Services allow large models to be widely used, and more tools, platforms and applications are needed to help companies implement scenarios. ——Tan Dai, President of Huoshan Engine 01. The large bean bag model makes its debut and is heavily used. Polishing the model effect is the most critical challenge for the implementation of AI. Tan Dai pointed out that only through extensive use can a good model be polished. Currently, the Doubao model processes 120 billion tokens of text and generates 30 million images every day. In order to help enterprises implement large-scale model scenarios, the beanbao large-scale model independently developed by ByteDance will be launched through the volcano
- AI 905 2024-06-05 19:59:21
-
- NVIDIA's new research: The context length is seriously false, and not many 32K performance is qualified
- Ruthlessly exposing the false standard phenomenon of "long context" large models - NVIDIA's new research found that 10 large models, including GPT-4, generate context lengths of 128k or even 1M. But after some testing, the new indicator "effective context" has shrunk seriously, and not many can reach 32K. The new benchmark is called RULER and includes a total of 13 tasks in four categories: retrieval, multi-hop tracking, aggregation, and question and answer. RULER defines the "effective context length", which is the maximum length at which the model can maintain the same performance as the Llama-7B baseline at 4K length. The research was rated "very insightful" by academics. After seeing this new research, many netizens also wanted to see the context length king players Claude and Gemini.
- AI 1039 2024-06-05 16:22:47
-
- The whole process of deploying yolov to iPhone or terminal practice
- The long-awaited detection classic has another wave of attacks - YOLOv5. Among them, YOLOv5 does not have complete files. The most important thing now is to figure out YOLOv4, which will benefit a lot in the field of target detection and can be highly improved in certain scenarios. Today we will analyze YOLOv4 for you. In the next issue, we will practice deploying YOLOv5 to Apple mobile phones or detect it in real time through the camera on the terminal! 1. Technology Review There are a large number of features that are considered to improve the accuracy of convolutional neural networks (CNN). Combinations of these features need to be practically tested on large datasets and the results theoretically validated. Some functions only operate on certain models, some questions only operate on certain models, or only small-scale problems.
- AI 514 2024-06-05 16:17:14
-
- A new chain of three-dimensional perception of embodied intelligence, TeleAI & Shanghai AI Lab proposed a multi-perspective fusion embodied model 'SAM-E'
- The AIxiv column is a column where this site publishes academic and technical content. In the past few years, the AIxiv column of this site has received more than 2,000 reports, covering top laboratories from major universities and companies around the world, effectively promoting academic exchanges and dissemination. If you have excellent work that you want to share, please feel free to contribute or contact us for reporting. Submission email: liyazhou@jiqizhixin.com; zhaoyunfeng@jiqizhixin.com When we pick up a mechanical watch, we will see the dial and hands from the front, the crown and bracelet from the side, and when we open the back of the watch, we will see the complex Gears and movements. Each perspective provides different information, which can be combined to understand the entire operation object.
- AI 535 2024-06-05 16:09:27
-
- The National Data Standardization Technical Committee officially approved the establishment of
- On the afternoon of May 24, 2024, the main forum of the 7th Digital China Construction Summit was held in Fuzhou. Liu Liehong, Secretary of the Party Leadership Group and Director of the National Data Administration, Tian Shihong, Member of the Party Leadership Group and Deputy Director of the State Administration for Market Regulation, Director of the National Standards Administration, and others attended the meeting. The meeting held in-depth discussions around the importance and strategic planning of digital China construction. Participants recognized that digital transformation has become a key driver of national development. At the same time, they emphasized the broad application of digital technology in the economy, society and governance. At the meeting, Tian Shihong read out the "Notice on Preparing for the Establishment of the National Data Standardization Technical Committee." The National Data Standardization Technical Committee will be responsible for data resources, data technology, data circulation, smart cities, digital transformation, etc.
- AI 490 2024-06-05 13:51:45
-
- Feature selection via reinforcement learning strategies
- Feature selection is a critical step in the process of building a machine learning model. Choosing good features for the model and the task we want to accomplish can improve performance. If we are dealing with high-dimensional data sets, then selecting features is particularly important. It enables the model to learn faster and better. The idea is to find the optimal number of features and the most meaningful features. In this article, we will introduce and implement a new feature selection via reinforcement learning strategy. We start by discussing reinforcement learning, specifically Markov decision processes. It is a very new method in the field of data science, especially suitable for feature selection. Then it introduces its implementation and how to install and use the python library (FSRLearning). Finally, a simple example is used to demonstrate this
- AI 465 2024-06-05 13:00:43
-
- A forward-looking perspective from HPE Aruba Networking: Building an AI-empowered secure converged network
- In the digital age, the network is not only a link connecting the world, but also a key force in promoting business and social progress. With the explosive growth of mobile devices, Internet of Things (IoT) and cloud services, users' dependence on the network has reached unprecedented levels. A network environment that can provide secure and efficient access anytime and anywhere has become a necessity for both individuals and enterprises. However, this growth in demand also brings its own set of challenges, particularly in terms of cybersecurity and the complexity of network architecture. In the latest Magic Quadrant for wired and wireless intervention released by Gartner this year, HPE Aruba Networking has been recognized again for its solutions in automated network orchestration, advanced software upgrade technology and integrated artificial intelligence (AI) features.
- AI 745 2024-06-05 11:12:47
-
- ECCV 2024 Workshop Multi-modal Understanding and Video Generation of Autonomous Driving Difficult Scenarios Call for Papers and Challenge is now open!
- Workshop home page: https://coda-dataset.github.io/w-coda2024/ Overview This workshop aims to explore the gap between the current state-of-the-art autonomous driving technology and comprehensive and reliable intelligent autonomous driving agents. In recent years, large multi-modal models (such as GPT-4V) have demonstrated unprecedented progress in multi-modal perception and understanding. Using MLLMs to deal with complex scenarios in autonomous driving, especially rare but critical hard-case scenarios, is an unsolved challenge. This workshop aims to promote innovative research in multi-modal large model perception and understanding, the application of advanced AIGC technology in autonomous driving systems, and end-to-end autonomous driving. Work
- AI 882 2024-06-04 20:47:35
-
- You can understand the principles of convolutional neural networks even with zero foundation! Super detailed!
- I believe that friends who love technology and have a strong interest in AI like the author must be familiar with convolutional neural networks, and must have been confused by such an "advanced" name for a long time. The author will enter the world of convolutional neural networks from scratch today ~ share it with everyone! Before we dive into convolutional neural networks, let’s take a look at how images work. Image Principle Images are represented in computers by numbers (0-255), and each number represents the brightness or color information of a pixel in the image. Among them: Black and white image: Each pixel has only one value, and this value varies between 0 (black) and 255 (white). Color image: Each pixel contains three values, the most common is the RGB (Red-Green-Blue) model, which is red, green and blue
- AI 390 2024-06-04 20:19:27
-
- 1.5 times beyond the diffraction limit, imaging conditions are 10 times lower, Tsinghua University and the Chinese Academy of Sciences use AI methods to improve microscope resolution
- Illustration: Long-term SR imaging of fast photosensitive biological processes via ZS-DeconvNet. (Source: Paper) Editor | Carrot skin computational super-resolution methods, including traditional analysis algorithms and deep learning models, have greatly improved optical microscopy. Among them, supervised deep neural networks have shown excellent performance, but due to the high dynamics of living cells, a large amount of high-quality training data is required, and obtaining these data is very laborious and impractical. In the latest study, researchers from Tsinghua University and the Chinese Academy of Sciences developed zero-shot deconvolution networks (ZS-DeconvNet) that can instantly improve the resolution of microscope images beyond the diffraction pole.
- AI 617 2024-06-04 19:26:15
-
- Focus on it! ! Analysis of two major algorithm frameworks for causal inference
- 1. The main tasks of the overall framework can be divided into three categories. The first is the discovery of causal structures, that is, identifying causal relationships between variables from the data. The second is the estimation of causal effects, that is, inferring from the data the degree of influence of one variable on another variable. It should be noted that this impact does not refer to relative nature, but to how the value or distribution of another variable changes when one variable is intervened. The last step is to correct for bias, because in many tasks, various factors may cause the distribution of development samples and application samples to be different. In this case, causal inference may help us correct for bias. These functions are suitable for a variety of scenarios, the most typical of which is decision-making scenarios. Through causal inference, we can understand how different users react to our decision-making behavior. Secondly, in industry
- AI 644 2024-06-04 16:45:02