当前位置:首页 > 报告详情

人工智能在金融领域的双刃剑.pdf

上传人: 杨*** 编号:878372 2025-08-16 58页 3.91MB

1、The Double-Edged Sword of AI in Finance:Secrets of Making AI WorkKeynote Talk1Usama M.FayyadInaugural Director of the Institute for Experiential AI-Northeastern UniversityProfessor of the Practice,Khoury College of Computer SciencesReWork AI in Finance Summit New York-April 18-19,2024What is Artific

2、ial Intelligence?The use of computers to“simulate”human intelligenceDefining“intelligence”is an open problem“Common Sense Reasoning”still an open problemWhat about Machine Learning?A subset of AI concerned with machines modifying/learning behaviors based on experience(inputs)-Training Data2The exces

3、sive hype lead to two AI Winters-Cut in funding,industry disillusionment,and practitioners avoid the fieldAI Winter 1-Mid 1970sAI Winter 2-Early 1990sSound like familiar hype?We are all going to be uselessJoblessBrainlessChina 2030 AI is the new Japanese 5thGenMajor hype in the 1980s AI was going to

4、 solve all problems and change the worldU.S.was afraid of Japan AI program 5th Gen.SystemsMachine Learning survived both AI wintersNot because we developed new/better ML algorithmsBut because we had a lot more dataHow does GenAI fit within AI,Machine Learning 5Artificial IntelligencePrograms with th

5、e ability tosimulate human intelligenceMachine LearningPrograms with the ability to learn without being explicitly programmedGenerative ModelsPrograms with the ability to learn how to generate new data that is similar to a given set of training dataSupervised Learning(Predictive AI)Applied MLUnsuper

6、vised LearningGen AIReinforcement LearningAfter“pre-training”,tune models to better align with human feedbackHuman feedback(Dall-E interpretation)Slides adapted from Primer Talk by Prof.Byron Wallace,Northeastern University-Generative AI Workshop:From the Classroom to the Economy-April 20236 6SECRET

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。
根据报告的内容,以下是全文关键点的概括: 1. **人工智能在金融中的应用**:人工智能在金融领域具有巨大潜力,包括信用评估、客户服务、欺诈检测和交易支持等。 2. **数据的重要性**:数据是人工智能工作的关键,尤其是高质量和细粒度的训练数据。 3. **人类干预的必要性**:人类在数据收集、模型训练和结果验证中扮演关键角色,以实现有效的AI合作。 4. **生成式AI的挑战**:生成式AI具有提高生产力和个性化体验的潜力,但也存在数据隐私、误信息和错误的风险。 5. **责任人工智能(RAI)**:实施RAI对于减少偏见、保护隐私和确保AI系统的透明度和可解释性至关重要。 6. **案例研究**:例如,Bloomberg GPT和Xfinance LLM展示了金融领域AI应用的实例。 7. **未来展望**:AI在知识经济中的地位日益重要,但需要合理的数据管理、人才培养和持续改进。
双刃剑还是机遇?" 数据与人才" "AI时代,如何打造竞争力?"
客服
商务合作
小程序
服务号
折叠