当前位置:首页 > 报告详情

AEC的未来:从副产品数据到以数据为中心的业务.pdf

上传人: 表表 编号:614350 2025-02-19 26页 19.77MB

1、April 2024Dr Tim WarkGlobal AI Lead,AECOM5th Ave NYC1900Where is the car?5th Ave NYC1913Where is the horse?Artificial Intelligence1956Artificial IntelligenceThe field of computer science that seeks to create intelligent machines that can replicate or exceed human intelligenceMachine Learning1997Mach

2、ine LearningSubset of AI that enables machines to learn from existing data and improve upon that data to make decisions or predictionsDeep Learning2017Deep LearningA machine learning technique in which layers of neural networks are used to process data and make decisionsGenerative AITransformer/Foun

3、dation Models2021Generative AICreate new written,visual,and auditory content given prompts or existing dataBrief history of AIAI:The growth of Large Language Models(LLMs)Source:twosigmaventuresGPT-4(1T)PaLM(540B)Megatron-Turing NLG(530B)Jurassic-1(178B)GPT-3(175B)Turing-NLG(17.2B)Megatron-LM(8.3B)T5

4、(11B)GPT-2(1.5B)BERT-Large(340M)ELMo(94M)MODEL SIZE(IN BILLIONS OF PARAMETERS)-5-The rise of cloud compute/GPUsSource: AI:Whats coming?Multimodality(e.g.video/images)ReasoningabilityReliability increaseIncreased customizability Shift from single tasks to broader agencyFundamental shiftsNatural langu

5、age becomes central to human-computer interfaceReasoning over unstructured dataThe rise and rise of data The rise and rise of data Bill Gates(Circa 1994)demonstrating the number of pieces of paper needed to store all the information on a CD-ROMTop 10 S&P 500 Companies(2010-2020)Source:chichaelmoh(ht

6、tps:/ rise and rise of data&computationalplatform companies*March 14,2024$3.08T*$2.64T*$1.74T*$1.83T*$2.27T*Approaches to managing enterprise dataTop downBottom upSource: data exampleFile Shares Technologies Cloud storage Windows File Servers(on-prem)Total Data Footprint:10s PB Data growth:TBs per d

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文概述了人工智能(AI)的发展历程,从1956年的概念提出,到1997年的机器学习,再到2017年的深度学习,直至2021年的生成式AI。文章指出,AI特别是生成式AI的兴起,将带来多模态、推理能力、可靠性和自定义性的提升,推动从单一任务到更广泛任务的转变。在建筑、工程和建设行业(AEC),生成式AI能够变革知识管理、加速内容创作、提升项目洞察力、支持商业策略、促进软件开发、优化设计流程和提升数字双生的预测能力。此外,预计生成式AI将促进多学科战略领导、系统思维者、数据科学家和咨询角色的增长。文章还提到,企业数据将成为宝贵资产,数据服务化的趋势将推动行业价值驱动因素的转变,如实时数据使用、客户需求定制和任务自动化。最后,建议AEC组织拥抱创新,建立安全准则,转变工作方式,构建信任文化,以适应AI带来的变革。
"AI在建筑行业中的未来趋势是什么?" "如何利用生成式AI提高设计和管理效率?" "AI在建筑行业中带来的挑战和机遇有哪些?"
客服
商务合作
小程序
服务号
折叠