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:
-
- BAIC Jihu releases Darwin 2.0 technology system, leading new energy vehicle technology to new heights
- On April 11, Tesla 2.0 technology was grandly released in Beijing. Jihu Motors once again leads the future with innovations in the field of smart travel. The Darwin 2.0 technology system is the result of Jihu Auto’s in-depth exploration and practice in the field of intelligent travel. It integrates seven major technology sectors, including intelligent driving, intelligent interconnection, intelligent safety and other aspects, to provide users with a comprehensive and intelligent travel experience. Through advanced sensors and algorithms, this technology system can achieve precise perception and intelligent decision-making of vehicle driving conditions, ensuring that users can enjoy a safe and comfortable driving experience in various traffic environments. At the same time, Darwin 2.0 technology also focuses on upgrading user experience. Through seamless connection with smart devices, it realizes the integration of in-vehicle systems with mobile phones and smart homes.
- AI 1031 2024-04-12 09:04:13
-
- Recommender systems based on causal inference: review and prospects
- The theme of this sharing is recommendation systems based on causal inference. We review past related work and propose future prospects in this direction. Why do we need to use causal inference techniques in recommender systems? Existing research work uses causal inference to solve three types of problems (see Gaoe et al.'s TOIS2023 paper Causal Inference in Recommender Systems: ASurvey and Future Directions): First, there are various biases (BIAS) in recommendation systems, and causal inference is an effective way to remove these Tools for bias. Recommender systems may face challenges in addressing data scarcity and the inability to accurately estimate causal effects. in order to solve
- AI 699 2024-04-12 09:01:07
-
- Long text cannot kill RAG: SQL+ vector drives large models and the new paradigm of big data, MyScale AI database is officially open source
- The combination of large models and AI databases has become a magic weapon for reducing costs and increasing efficiency for large models and making big data truly intelligent. The wave of large models (LLM) has been surging for more than a year, especially models represented by GPT-4, Gemini-1.5, Claude-3, etc., which have become a well-deserved hot spot. On the LLM track, some research focuses on increasing model parameters, and some are crazy about multi-modality... Among them, LLM's ability to process context length has become an important indicator for evaluating models. A stronger context means that the model Have stronger retrieval performance. For example, the ability of some models to process up to 1 million tokens in one go has led many researchers to think about RAG (R
- AI 1239 2024-04-12 08:04:24
-
- Face Wall Intelligence completed a new round of financing of several hundred million yuan and continued its journey towards efficient large models for AGI
- Recently, Wall-facing Intelligence completed a new round of financing of several hundred million yuan, led by Primavera Ventures and Huawei Hubble, and followed by the Beijing Artificial Intelligence Industry Investment Fund, with Zhihu continuing to invest as a strategic shareholder. After the completion of this round of financing, Face Wall Intelligence will further promote the introduction of outstanding talents, strengthen the basic computing power and data foundation for large model deployment, continue to lead the "efficient large model" route, promote efficient training of large models, and quickly implement applications. Light Source Capital served as the exclusive financial advisor for this round. Based on a solid foundation of original AI technology, Wall-Facing Intelligence is one of the largest large model teams in the world that is at the forefront of exploring "efficient large models". It has now completed large model full-stack technology that implements efficient training, efficient implementation and efficient reasoning. Production line layout. The core R&D team was born out of Tsinghua NL
- AI 1053 2024-04-11 21:22:01
-
- Exploring the boundaries of agents: AgentQuest, a modular benchmark framework for comprehensively measuring and improving the performance of large language model agents
- Based on the continuous optimization of large models, LLM agents - these powerful algorithmic entities have shown the potential to solve complex multi-step reasoning tasks. From natural language processing to deep learning, LLM agents are gradually becoming the focus of research and industry. They can not only understand and generate human language, but also formulate strategies, perform tasks in diverse environments, and even use API calls and coding to Build solutions. In this context, the introduction of the AgentQuest framework is a milestone. It not only provides a modular benchmarking platform for the evaluation and advancement of LLM agents, but also provides researchers with a Powerful tools to track and improve the performance of these agents at a more granular level
- AI 1102 2024-04-11 20:52:21
-
- Multiple SOTAs! OV-Uni3DETR: Improving the generalizability of 3D detection across categories, scenes and modalities (Tsinghua & HKU)
- This paper discusses the field of 3D object detection, especially 3D object detection for Open-Vocabulary. In traditional 3D object detection tasks, systems need to predict the localization of 3D bounding boxes and semantic category labels for objects in real scenes, which usually relies on point clouds or RGB images. Although 2D object detection technology performs well due to its ubiquity and speed, relevant research shows that the development of 3D universal detection lags behind in comparison. Currently, most 3D object detection methods still rely on fully supervised learning and are limited by fully annotated data under specific input modes, and can only recognize categories that emerge during training, whether in indoor or outdoor scenes. This paper points out that the challenges facing 3D universal target detection are mainly
- AI 376 2024-04-11 19:46:18
-
- AI security company TrojAI receives additional seed funding
- TrojAI, a Canadian AI security solutions provider, announced this week that it has received $5.75 million in additional seed funding. The enterprise AI security platform provided by TrojAI helps customers protect AI models and applications from risks and attacks. Its platform can test AI models before deployment and protect applications from issues such as sensitive data leakage, helping enterprises comply with benchmarks such as the OWASPAI framework and privacy regulations. Its main business modules are as follows: AI model risk detection: The TrojAI platform can be integrated with AI and MLOps workflows to automatically penetrate and scan AI models before production to identify potential risks and vulnerabilities, such as backdoors, data leaks and bias. AI application protection: TrojAI platform can protect A
- AI 815 2024-04-11 19:43:17
-
- By 2028, the smart building AI market is expected to reach US$6.48 billion
- This new research, based on analysis of the artificial intelligence (AI) market, looks at the progress being made in AI’s broad capabilities and its specialized applications that are making the built environment smarter, more sustainable, and more responsive. This is the first in a two-part series, with the second on the AI market landscape to be released later this year. This report explores where we are on the journey towards “truly cognitive architecture”. Today’s commercial construction technology is moving from rules-based analytics to artificial intelligence predictive machine learning models, but adoption rates are still at middling levels. The scope of real-world deployments remains narrow, driven primarily by more well-understood use cases in energy optimization, space utilization, and security. Challenges and obstacles hindering the application of artificial intelligence
- AI 982 2024-04-11 19:31:16
-
- Ant Group's CodeFuse launches the 'Picture to Generate Code” function, and more than 50% of programmers use AI to write code
- On April 11, CodeFuse, Ant Group’s self-developed intelligent R&D platform, launched a new function called “Picture Generating Code”, which allows developers to generate code with one click using product design drawings, greatly improving the development efficiency of front-end pages. The relevant functions are currently in internal testing. Like many Internet companies, Ant Group is fully promoting AI programming. More than 50% of engineers use CodeFuse to support daily research and development work. 10% of the codes submitted by these engineers are generated by AI. Gartner pointed out in the top ten strategic technology trends released in 2024: By 2028, 75% of enterprise software engineers will use AI programming assistants. CodeFuse is an exploration attempt under this trend. According to reports, CodeFus
- AI 504 2024-04-11 18:52:22
-
- The 'Devin AI era' of programming, the joys and worries of software developers
- Author | Compiled by Keith Pitt | Produced by Yifeng | 51CTO Technology Stack (WeChat ID: blog51cto) The author of this article, Keith Pitt, is the founder and CEO of Buildkite, a software development company. In 2013, he founded the company with another software engineer, Tim Lucas, to provide a continuous integration and continuous delivery (CI/CD) platform for the technology industry. It recently received support from OneVentures and AirTree. Co-led $21 million in Series B financing. A 20-year programming veteran and CEO of a company that serves software developers, Keith Pitt (K
- AI 1157 2024-04-11 17:10:12
-
- Six trends that will have a major impact on today's corporate IT market
- Everyone is talking about AI and it is pointed out that many companies are already integrating AI into their business. "It's already built in, or is being built into existing SaaS platforms at the largest providers." However, with the arrival of AI comes a certain level of fear and apprehension, Fox said. There are many problems. "What would a fully AI-enabled company look like? Will it have the same workforce, in the same locations?" Those aren't questions that need to be answered today, she said, but they do need to be considered. This and several other emerging trends may be about to reshape the IT and business world. That's why Asana CEO Saket Srivastava believes there has never been a better time to become a CIO. Despite uncertainty about AI and the future of work
- AI 901 2024-04-11 17:07:01
-
- Efficiency increased by 16 times! VRSO: 3D annotation of purely visual static objects, opening up the data closed loop!
- Annotation of static object detection (SOD), including traffic lights, guide signs and traffic cones, most algorithms are data-driven deep neural networks and require a large amount of training data. Current practice typically involves manual annotation of a large number of training samples on LiDAR-scanned point cloud data to fix long-tail cases. Manual annotation has difficulty capturing the variability and complexity of real scenes and often fails to account for occlusions, different lighting conditions, and diverse viewing angles (yellow arrows in Figure 1). The entire process has long links, is extremely time-consuming, error-prone, and costly (Figure 2). Therefore, companies are currently looking for automatic labeling solutions, especially based on pure vision. After all, not every car has lidar. VRSO+ is a visual-based labeling oriented for static object labeling.
- AI 670 2024-04-11 16:16:20
-
- How can artificial intelligence make computing easier?
- Artificial intelligence (AI) and machine learning (ML) are becoming increasingly common in our daily lives, but we often don’t realize it. These technologies simplify every aspect of computing, making it more efficient, accessible, and user-friendly. Simplification and Integration of Artificial Intelligence The "Simplicity and Power" (SP) theory of intelligence proposed by Gerry and Wolf is a research project dedicated to the development of artificial intelligence. Rather than focusing on a single area of AI, such as reasoning or computer vision, SP theory aims to develop a framework applicable to multiple AI disciplines. By simplifying and integrating observations and concepts from personal intelligence, mainstream computing, mathematics, human learning, perception and cognition, SP theory successfully creates a unified framework to express various kinds of knowledge and intelligence.
- AI 933 2024-04-11 15:55:08
-
- Stimulate the spatial reasoning ability of large language models: thinking visualization tips
- Large language models (LLMs) demonstrate impressive performance in language understanding and various reasoning tasks. However, their role in spatial reasoning, a key aspect of human cognition, remains understudied. Humans have the ability to create mental images of unseen objects and actions through a process known as the mind's eye, making it possible to imagine the unseen world. Inspired by this cognitive ability, researchers proposed "Visualization of Thought" (VoT). VoT aims to guide the spatial reasoning of LLMs by visualizing their reasoning signs, thereby guiding subsequent reasoning steps. Researchers apply VoT to multi-hop spatial reasoning tasks, including natural language navigation, vision
- AI 1107 2024-04-11 15:10:17
-
- AI-driven workspaces: a blessing or a curse?
- AI-driven workspaces: a blessing or a curse? Artificial intelligence (AI) has made significant progress in various industries, and its integration into the workspace is inevitable. Artificial intelligence-powered workspaces promise to revolutionize the way we increase productivity and collaboration, while raising concerns about privacy, security and the future of work. This article explores the pros and cons of AI-driven workspaces. Benefits of AI-powered workspaces AI-powered workspaces offer many benefits, including increased productivity, improved collaboration, and enhanced user experience. By automating routine tasks, AI can help employees save time and focus on more strategic and creative work. For example, GPTforDocs, DialpadAI
- AI 537 2024-04-11 15:04:30