Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

AI for Retail Banking


Published on

AI/Machine Learning Cases for Retail Banking

Published in: Data & Analytics
  • Be the first to comment

AI for Retail Banking

  1. 1. AI for Retail Banking Dmitry Petukhov Microsoft MVP, ML/DS Preacher @ OpenWay Moscow Cognitive Computing Community #m3community
  2. 2. Customer Segmentation Financial Markets & etc. Retail Banking Insurance Real-time Batch processingDuration Market Assets Price Prediction Social Network Analysis Fraud Detection Risk Analysis Compliance & Regulatory Reporting Advertising Campaign Optimization News Analysis Customer Loyalty & Marketing Improving operational efficiencies Credit Scoring Brand Sentiment Analysis Personalized Product Offering AI for Retail Banking: Use Cases in Finance
  3. 3. Personalized Product Offering Real-timeBatch Processing Processing Speed Log(Volume) Pbytes Tbytes Gbytes Structured data Semi-structured Unstructured Customer Loyalty & Marketing Fraud Detection & Security Credit Scoring Compliance & Regulatory Reporting Operational Efficiencies Customer Segmentation Voice identity, Chat-bots, Person Financial Manager AI for Retail Banking: Use Cases in Retail Banking
  4. 4. AI for Retail Banking: Use Cases in Retail Banking Алгоритмы машинного обучения: C – классификация (Classification); CA – кластерный анализ (Cluster Analysis); LSA – латентно-семантический анализ (Latent Semantic Analysis); AD – обнаружение аномалий (Anomaly Detection); CF – коллаборативная фильтрация (Collaborative Filtering). Источники данных: Transactions Log – лог финансовых транзакций; Banking/Merchant CRM Data – CRM-профили клиента/мерчанта; Web-applications Log – логи интернет- и мобильного банков; External Services – внешние DMP, такие как НБКИ; Support Service Data – данные отдела клиентской поддержки; Social Network Data – социальные сети.
  5. 5. Клиент (web-браузер) Мерчант (интернет-магазин) Электронная платежная система Банк-эквайер мерчанта Банк-эмитент Международная платежная система 1 2 9 8 4 3 7 4 6 5 Real time Not real time AI for Retail Banking: Antifraud in E-commerce
  6. 6. AI for Retail Banking: Antifraud Statistics Компания Источник Показатель / результат Яндекс.Деньги Выступление фрод-аналитика Яндекс.Деньги, конференция Antifraud Russia 2015 Карточное мошенничество России за 2015 год - 3,5 млрд. руб. Антифрод-система Яндекс.Деньги, основанная на алгоритмах ML, отлавливает >90% фродовых транзакций PayOnline Отчет «Мошенничество в Рунете» CNP-мошенничество в России за 2015 год - 1,2 млрд. руб. (+45%) Сбербанк Выступление Германа Грефа, годовое собрание акционеров Сбербанка Анализ поведенческой активности держателя карт, основанный на алгоритмах ML, останавливает фрод на 150-200 млн. руб. в неделю Assist Выступление «Data Science для обеспечения безопасности платежей», конференция Платежные инновации и... Снижение уровня отклоненных по 3DS транзакций с 18,9% до 1,4% за счет интеллектуального анализа клиентских данных Accertify, ACI Worldwide, Agnitio, Ayasdi, BAE Systems Applied Intelligence, BioCatch, CA Technologies, Contact Solutions, CustomerXPs, CyberSource, Digital Resolve, Easy Solutions, Experian (41st Parameter), F5 (Versafe), Feedzai, Fox-IT, GBGroup, Guardian Analytics... and 25 more Source: Gartner Inc., 2015
  7. 7. External Services: DMP-data, geolocation, etc. Customer Support Service Data Black/white Lists of Plastic Cards, Merchants, IP-hosts, etc. Number of customer grows fast… Number of operations grows even faster… Transactions Log with request information Banking CRM Data Merchant CRM Data Web-clicks Stream Web/Mobile-applications & Backend Services Log Data for Model Join data Pain AI for Retail Banking: Antifraud in E-commerce
  8. 8. Quality AI for Retail Banking: Antifraud in E-commerce
  9. 9. Storage Resource Management ML Framework Execution Engine Local OS Local Disc PythonRuntime YetAnother Runtime scikit learn HDFS YARN MapReduce Mahout HDFS / S3 YARN / Apache Mesos Spark MLlib HDFS / S3 YARN / Apache Mesos Python / R on Spark Python / R tools Spark Local PC Hybrid Model Cluster (on-premises/on-demand) some library Low HighCost of deployment/ownership Distributed FS Dark Magic… ML as a Service Python / R tools AI for Retail Banking: Antifraud in E-commerce
  10. 10. AI for Retail Banking: Innovations It is Future Deep Learning Identity and access management (IAM) services Biometric methods: voice, fingers, eyes, heartbeats(!) Personal financial manager Intelligent personal assistant Income/withdraw extrapolation (+linear regression) Personalized product offering (+logistic regression) Customer Support Voice recognition: customer identity, emotions, conversation essence (!) Chat-bots
  11. 11. FinTech Startups FinTech Incubators & Accelerators AlfaCamp Barclays Accelerator MasterCard Start Path Visa Europe Collab QIWI Universe 2.0 InspirAsia (Life.SREDA) Future Fintech to be continued… Researchers & Enthusiasts Competitions & Hackathons Sberbank Alfabank Tinkoff Otkritie to be continued… AI for Retail Banking: Opportunities Time
  12. 12. habr blog github AI for Retail Banking: Practices ML in Finance – Present and Future Machine learning for financial prediction
  13. 13. © 2016 Dmitry Petukhov. All rights reserved. Microsoft and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. Thank you!
  14. 14. Q&A Now or later (by e-mail) Stay in Touch! Facebook: @code.zombi Habr: @codezombie All contacts on Download this presentation from or