人工智能的工程数据可靠性.pdf

1、Engineering Data Reliability for AIQuick Poll How many of you are.Considering AI projects within your enterprise?Actively implementing AI projects within your enterprise?Are focusing on data reliability for AI?123Enterprise AI AdoptionActively DeployingExploring or ExperimentingNot Exploring42%40%18

2、%Improving Operational EfficiencyEnhancing Customer ExperienceGaining Competitive AdvantageSource:https:/ Enterprise DriversEnterprise AI AdoptionTransformational Opportunity of AIAI spending surged to$13.8 billion in 2024,a 6X increase from 2023AI could contribute$15.7 trillion to the global econom

3、y by 2030By 2030,AI is expected to boost North Americas GDP by 14.5%But there is this.80%AI projects likely to fail 2XFailure rate of IT projects30%AI projects likely to be abandoned after proof of concept by the end of 2025Sources1.80%of AI Projects Fail-Why?And What Can We Do About It?-IHL Group2.

4、The Surprising Reason Most AI Projects Fail And How to Avoid It at Your Enterprise|Informatica3.Why AI projects fail and how to save yoursWHY?DATA READINESS FOR AI6.A KEY FACTOR3Age Old AxiomHidden Cost of Hidden Cost of Data DefectsData Defects$15M$15MDifficult to UndoDownstream ImpactFinancial&Com

5、pliance RisksHigh Cost To FixImpacts BusinessContinuityannually on an average for enterprisesannually on an average for enterprisesSource:(4)How Much is Poor Quality Data Really Costing your Business?|LinkedInWHERE CURRENT ENTERPRISE DATA QUALITY PROGRAMS FOCUS!Source Data VariabilityDegraded model

6、accuracyCommon Data Gaps&AI FailuresDATAData&Schema DriftPipeline FailuresMissed Business RulesLack of metrics and observationAnomaliesBreaks in features and trainingData loss and stale modelsMisaligned insights&recosBusines risk introductionOutliers and biases“Id rather get a root canal than fix my