AISummit successfully opened: Highlights of the first day
In midsummer and August, the sun is like fire and the vegetation is lush, everything shows the wild and poetic vitality. Summer is a season of exploration, growth, and innovation. In this season that belongs to practitioners, 51CTO brought an AI event with the theme of "Drive·Innovation·Digital Intelligence".
AI technology was born less than a century ago. After several ups and downs, it has ushered in a golden period of comprehensive development and implementation in the past 20 years. What are the current cutting-edge technological achievements and practical innovation breakthroughs in the field of AI? How do you view the next decade of AI? This is an issue lingering in the minds of many technicians.
On August 6, the AISummit Global Artificial Intelligence Technology Conference opened as scheduled as an online live broadcast. On the first day, nearly a hundred experts, scholars, technical experts, and management elites gathered together to discuss the wave of "digital intelligence" in the era of artificial intelligence with tens of thousands of participants.
Looking from a high position, interpreting the current situation and trends of AI
In the opening speech, Mr. Cui Kang, Vice President and Editor-in-Chief of 51CTO, expressed his expectations for the conference as the main planner of this conference. He believes that with the development of artificial intelligence, it is no longer like a segmented technology in the IT field, but more like a general technology. Although its development is improving all the way, it still needs to face many current contradictions. This conference hopes to provide some enlightening clues for these contradictions and find some solutions. At the same time, I also hope that “this conference can tell participants that AI technology and AI developers are making the world a better place.”
In the main forum in the morning, six important guests interpreted the current status and trends of artificial intelligence technology from the perspective of technology leaders, the research findings that practitioners need to pay attention to, and the impact of cutting-edge technology in the industry.
Dou Zhicheng, deputy dean of the Hillhouse School of Artificial Intelligence at Renmin University of China, shared his prospects for the next generation of intelligent search technology. He interpreted the development trends and core features of the new generation of intelligent search technology, and also made a detailed analysis of technologies such as interactive, multi-modal, interpretable search, and large model-centered de-indexed search.
Tian Yuandong, researcher and senior manager of Meta/Facebook Artificial Intelligence Research Institute, focused on analyzing the opportunities and challenges of decision-making in real-world scenarios. Currently, how to use deep neural networks to process structured data and find neural network solutions that replace human heuristic strategies for some discrete optimization problems is still an unsolved problem. Tian Yuandong introduced how to use reinforcement learning and search methods with neural networks to find heuristic algorithms for complex optimization problems.
Pan Qinghua, deputy director of iFlytek Research Institute of HKUST, gave his own judgment on the development trend of artificial intelligence technology using the industry implementation of AI as an entry point. In-depth answers were given around how to form an interactive method that embodies emotions and combines reality and reality, how to create a complex intelligent system that integrates multi-modal modes and integrates software and hardware, and how to develop better solutions for the industry.
Zhou Zongwei, a senior engineer and senior manager at Google and the leader of the MLPerf team, summarized ten points of understanding that affect the design of Google's artificial intelligence chips from multiple perspectives based on the explosive growth in computing power demand in recent years. In terms of his views on the new generation The impact of the architecture and design of the quasi-structured chip was sorted out, and the collaborative evolution of Google's chips and software was interpreted.
Xiang Liang, head of the Volcano Engine machine learning system, mainly introduced the technological evolution of large-scale machine learning computing power. In the keynote speech, he analyzed the technical difficulties and diminishing returns encountered in the implementation of large models, conducted an in-depth analysis of the promotion and impact between the computing power and machine learning industries, and also analyzed the future development trends of computing power. Predictions were made.
Liu Yi, head of Microsoft's Bing advertising text generation algorithm, took Microsoft's Bing DeepGen project as an example to explain the issues of diversified search advertising text generation and online real-time matching. The advertising text generation technology is introduced in detail, including system brief introduction, basic and diversified generation models, algorithm model, model characteristics and commercial impact of real-time matching of advertising online text.
Multi-dimensional interpretation, focus on the highlights of the sub-forums
The four sub-forums held in the afternoon were also exciting and climaxed one after another. Its content covers four major topics: "AI-driven search and recommendation", "MLOps best practices", "machine learning performance optimization road", and "computer vision application and innovation" and 16 subdivided topic directions.
AI-driven search and recommendation
The wave of digital transformation has given rise to a new evolution of search recommendation technology, and recommendation technology has also ushered in a stage of deep integration with intelligent algorithms. In the "AI-driven search and recommendation" sub-forum, senior experts in the field of search and recommendation shared forward-looking thinking in the field of intelligent search and recommendation from the perspective of business practice.
Wang Liang, senior technical expert at Alimama and head of external advertising technology, explained the concept of federated learning and its application in Alibaba advertising; Ma Jianqiang, senior researcher at Tencent and head of online video knowledge graph, explained Tencent’s With the vertical search of video search as the background, it focused on the main technical scenarios, algorithm architecture and progress of video search, short video vector recall, the application of long video IP knowledge graph, etc.; Chen Sheng, head of Meituan search ranking, said "Meituan search" With the theme of "Construction and Practice of Sorting Platform", it explained in detail the technical architecture of Meituan Search, the construction of the sorting platform, and the optimization of the sorting algorithm; Kuaishou senior recommendation algorithm expert Zang Xiaoxue brought Kuaishou's experience in causal inference and Latest research on graph neural network algorithms.
MLOps Best Practices
In the large-scale application of artificial intelligence in enterprises, there are challenges such as long R&D rollout cycles and difficulty in matching data and models. MLOps came into being. In the "MLOps Best Practices" sub-forum, experts discussed the practical effects and effectiveness of MLOps around hot topics such as the R&D operation and maintenance cycle, continuous training and continuous monitoring, model version and lineage, online and offline data consistency, and efficient data supply. Cutting edge trends.
Tan Zhongyi, Vice Chairman of TOC of Open Atomic Foundation and member of LF AI & Data TAC, introduced the concept, positioning, main content, common projects of MLOps, and the criteria for evaluating the ability and level of MLOps of an AI team; Chapter 1 Lu Mian, Four Paradigm system architect and head of OpenMLDB R&D, focused on the open source machine learning database OpenMLDB and analyzed how it can achieve the goal of immediately launching machine learning feature development and how to ensure the correctness and efficiency of feature calculation; NetEase Cloud Music Artificial Intelligence Researcher and Technical Director Wu Guanlin started from the background of cloud music business and explained the real-time implementation plan of the model, and the thinking and practice of further radiating to more scenarios by combining FeatureStore; Huang Bingkey, Deputy Director of the Big Data and Artificial Intelligence Laboratory of the Software Development Center of the Industrial and Commercial Bank of China This article introduces the MLOps practice of ICBC, covering the construction process and technical practice of the full life cycle management system of model development, model delivery, model management, and model iterative operation.
The Road to Machine Learning Performance Optimization
In the field of machine learning, enterprises have continuously increasing performance requirements for algorithms: How to ensure the stability of the system? How to use engineering means to solve the problem of insufficient samples and real-time model? How to improve ease of use? In the "Road to Machine Learning Performance Optimization" sub-forum, experts shared their opinions on this.
Yang Yang, chief engineer of Didi, explained the innovative application of personalized recommendations in data operations based on the pain points of traditional data operations, and looked forward to the technologies and businesses that need to be improved in future data operations; 100 people from Tongji University Program Distinguished Researcher and Doctoral Supervisor Wang Haofen explained the key technologies and commonly used data sets involved in multi-strategy question answering of text knowledge and two forms of multi-modal question answering; Cai Qingpeng, a senior algorithm expert at Kuaishou, used reinforcement learning in Kuaishou short video recommendation The system's technical implementation is used as a case to explain Kuaishou's practical experience in online parameter search based on reinforcement learning, two-stage constrained reinforcement learning algorithm, and how to optimize and achieve APP activity; NetEase Cloud Music Algorithm Platform R&D expert Huang Bin introduced NetEase Cloud Music Online The practice and thinking of prediction systems include how to build a high-performance and easy-to-use prediction system, and how to solve problems such as real-time samples and models through engineering means.
Computer Vision Application and Innovation
Computer vision (CV), as the pioneer of AI technology, is the foundation of many innovative key technologies. In the "Computer Vision Application and Innovation" sub-forum, experts in the CV field conducted in-depth analysis on multiple scenarios such as video quality monitoring, intelligent video creation, and autonomous driving.
Li Jing, director of Alibaba Youku Technology Center and head of Moku Lab, proposed using AI to improve the short video creation process in response to existing problems in the field of short video creation, and shared the exploration and development of Youku's AI video intelligent production system. Practice; SSIMWAVE co-founder and chief researcher Zeng Kai explained how to solve possible problems in the end-to-end video quality monitoring system and the AI-based objective video quality evaluation algorithm; Tang Dong, head of Ant Group's large security image similarity traceability algorithm Qi used Ant Security Technology's "Siyuan" similar traceability engine as a practical case to conduct a technical dismantling of how to identify and trace in open scenarios; if Ma Zhiguo, an expert in smart mobility perception technology, analyzed the lidar solution in autonomous driving and explained the automatic Perception technology involved in driving, and an in-depth analysis of the relationship between data and autonomous driving.
More exciting, please stay tuned
Taking this AISummit conference as an opportunity, 51CTO will continue to explore and try with ecological partners in the future to build an artificial intelligence for the majority of technical personnel A platform for in-depth communication and sharing in the field. On August 7, the conference will also usher in special sessions on intelligent voice, smart finance, and the metaverse. Friends who follow this conference can look forward to the exciting continuation.
The above is the detailed content of AISummit successfully opened: Highlights of the first day. For more information, please follow other related articles on the PHP Chinese website!
Hot AI Tools
Undresser.AI Undress
AI-powered app for creating realistic nude photos
AI Clothes Remover
Online AI tool for removing clothes from photos.
Undress AI Tool
Undress images for free
Clothoff.io
AI clothes remover
Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!
Hot Article
Hot Tools
Notepad++7.3.1
Easy-to-use and free code editor
SublimeText3 Chinese version
Chinese version, very easy to use
Zend Studio 13.0.1
Powerful PHP integrated development environment
Dreamweaver CS6
Visual web development tools
SublimeText3 Mac version
God-level code editing software (SublimeText3)
Hot Topics
Bytedance Cutting launches SVIP super membership: 499 yuan for continuous annual subscription, providing a variety of AI functions
Jun 28, 2024 am 03:51 AM
This site reported on June 27 that Jianying is a video editing software developed by FaceMeng Technology, a subsidiary of ByteDance. It relies on the Douyin platform and basically produces short video content for users of the platform. It is compatible with iOS, Android, and Windows. , MacOS and other operating systems. Jianying officially announced the upgrade of its membership system and launched a new SVIP, which includes a variety of AI black technologies, such as intelligent translation, intelligent highlighting, intelligent packaging, digital human synthesis, etc. In terms of price, the monthly fee for clipping SVIP is 79 yuan, the annual fee is 599 yuan (note on this site: equivalent to 49.9 yuan per month), the continuous monthly subscription is 59 yuan per month, and the continuous annual subscription is 499 yuan per year (equivalent to 41.6 yuan per month) . In addition, the cut official also stated that in order to improve the user experience, those who have subscribed to the original VIP
Context-augmented AI coding assistant using Rag and Sem-Rag
Jun 10, 2024 am 11:08 AM
Improve developer productivity, efficiency, and accuracy by incorporating retrieval-enhanced generation and semantic memory into AI coding assistants. Translated from EnhancingAICodingAssistantswithContextUsingRAGandSEM-RAG, author JanakiramMSV. While basic AI programming assistants are naturally helpful, they often fail to provide the most relevant and correct code suggestions because they rely on a general understanding of the software language and the most common patterns of writing software. The code generated by these coding assistants is suitable for solving the problems they are responsible for solving, but often does not conform to the coding standards, conventions and styles of the individual teams. This often results in suggestions that need to be modified or refined in order for the code to be accepted into the application
Seven Cool GenAI & LLM Technical Interview Questions
Jun 07, 2024 am 10:06 AM
To learn more about AIGC, please visit: 51CTOAI.x Community https://www.51cto.com/aigc/Translator|Jingyan Reviewer|Chonglou is different from the traditional question bank that can be seen everywhere on the Internet. These questions It requires thinking outside the box. Large Language Models (LLMs) are increasingly important in the fields of data science, generative artificial intelligence (GenAI), and artificial intelligence. These complex algorithms enhance human skills and drive efficiency and innovation in many industries, becoming the key for companies to remain competitive. LLM has a wide range of applications. It can be used in fields such as natural language processing, text generation, speech recognition and recommendation systems. By learning from large amounts of data, LLM is able to generate text
Can fine-tuning really allow LLM to learn new things: introducing new knowledge may make the model produce more hallucinations
Jun 11, 2024 pm 03:57 PM
Large Language Models (LLMs) are trained on huge text databases, where they acquire large amounts of real-world knowledge. This knowledge is embedded into their parameters and can then be used when needed. The knowledge of these models is "reified" at the end of training. At the end of pre-training, the model actually stops learning. Align or fine-tune the model to learn how to leverage this knowledge and respond more naturally to user questions. But sometimes model knowledge is not enough, and although the model can access external content through RAG, it is considered beneficial to adapt the model to new domains through fine-tuning. This fine-tuning is performed using input from human annotators or other LLM creations, where the model encounters additional real-world knowledge and integrates it
Five schools of machine learning you don't know about
Jun 05, 2024 pm 08:51 PM
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
To provide a new scientific and complex question answering benchmark and evaluation system for large models, UNSW, Argonne, University of Chicago and other institutions jointly launched the SciQAG framework
Jul 25, 2024 am 06:42 AM
Editor |ScienceAI Question Answering (QA) data set plays a vital role in promoting natural language processing (NLP) research. High-quality QA data sets can not only be used to fine-tune models, but also effectively evaluate the capabilities of large language models (LLM), especially the ability to understand and reason about scientific knowledge. Although there are currently many scientific QA data sets covering medicine, chemistry, biology and other fields, these data sets still have some shortcomings. First, the data form is relatively simple, most of which are multiple-choice questions. They are easy to evaluate, but limit the model's answer selection range and cannot fully test the model's ability to answer scientific questions. In contrast, open-ended Q&A
SOTA performance, Xiamen multi-modal protein-ligand affinity prediction AI method, combines molecular surface information for the first time
Jul 17, 2024 pm 06:37 PM
Editor | KX In the field of drug research and development, accurately and effectively predicting the binding affinity of proteins and ligands is crucial for drug screening and optimization. However, current studies do not take into account the important role of molecular surface information in protein-ligand interactions. Based on this, researchers from Xiamen University proposed a novel multi-modal feature extraction (MFE) framework, which for the first time combines information on protein surface, 3D structure and sequence, and uses a cross-attention mechanism to compare different modalities. feature alignment. Experimental results demonstrate that this method achieves state-of-the-art performance in predicting protein-ligand binding affinities. Furthermore, ablation studies demonstrate the effectiveness and necessity of protein surface information and multimodal feature alignment within this framework. Related research begins with "S
SK Hynix will display new AI-related products on August 6: 12-layer HBM3E, 321-high NAND, etc.
Aug 01, 2024 pm 09:40 PM
According to news from this site on August 1, SK Hynix released a blog post today (August 1), announcing that it will attend the Global Semiconductor Memory Summit FMS2024 to be held in Santa Clara, California, USA from August 6 to 8, showcasing many new technologies. generation product. Introduction to the Future Memory and Storage Summit (FutureMemoryandStorage), formerly the Flash Memory Summit (FlashMemorySummit) mainly for NAND suppliers, in the context of increasing attention to artificial intelligence technology, this year was renamed the Future Memory and Storage Summit (FutureMemoryandStorage) to invite DRAM and storage vendors and many more players. New product SK hynix launched last year
