How does Deep Learning can help banking players? Heuritech at Groupe BPCE conference

CeE1zDcWAAEJxZ4

Groupe BPCE conference « Opening up on Digital Transformation, Machine Learning & Big Data: What tomorrow’s tech has to offer for BPCE today » was the opportunity for Heuritech to present some Deep Learning business use cases and perspectives of application.

Tony Pinville and Charles Thurat, both PhDs in Artificial Intelligence, presented 3 missions that Heuritech performed and how Deep Learning was put to task to answer for very concrete use cases :

  • Use case n° 1: helping a major banking and credit actor in better knowing its customers and predicting its needs.
  • Use case n° 2 : extract value from unstructured databases for decision support to a financial management division.
  • Use case n° 3 : real-time web media monitoring for brand analysis, competitive intelligence and market research.

Then they developed from these examples to provide new perspectives directly linked to these use cases, or more distantly related but entirely relevant future uses for this technology :

  1. General-purpose uses for businesses – The cross-language and technical vocabulary learning abilities offered by Machine Learning, as well as its robustness in extracting and analysis contents from unstructured databases allow the following uses:
    • Intelligent search engine for enterprise search (generalisation over use case 2).
    • Automated and precise indexation and categorisation of documents (cross-over between uses cases 1 and 2).
    • Automated multilingual dictionary creation in order to clarify technical terms in technical and reglementary texts, with automated detection of redundancies, discrepancies and contradictions.
  2. Customer risk prediction by cross-analysis of 1st and 3nd party data – The solutions proposed for the credit actor use case can be expanded to other problematics involving the profiling of clients and/or customers and the detection of their habit patterns:
    • Automated profiling, categorisation and segmentation of clients and prospects.
    • Automated behavior analysis and pattern detection to check unusual activities (fraud detection).
    • Automated verification of client databases conformity to privacy regulations (CNIL).
    • Automated mapping of connections between clients to detect redundancies and evaluate conformity of operations and services provided to multiples aliases of the same person.
  3. Conformity risk and audit analysis and prediction:
    • Identification of risks in procedures conformity for control and audit activities.
    • Automated evaluation of the needs/ resources adequation  for control plans, and evaluation of their application.
    • Automated evaluation and cartography of risks, evaluation of operational procedures adequation to said risks.
    • Prediction of the evolution of procedures compliance to changing regulations
  4. Decision support by multichannel analysis of 3rd party data:
    • Automated benchmark analysis of market and sector  trends and by aggregating and analysing multiple online sources.
    • Automated strategic / competitive intelligence
  5. In the more distant future:
    • Fusion of image and text analysis for the automated evaluation of the value of any good through cross-analysis of multiple 3rd party data channels.

And thanks again to Groupe BPCE for this conference.

Laisser un commentaire

Entrez vos coordonnées ci-dessous ou cliquez sur une icône pour vous connecter:

Logo WordPress.com

Vous commentez à l'aide de votre compte WordPress.com. Déconnexion / Changer )

Image Twitter

Vous commentez à l'aide de votre compte Twitter. Déconnexion / Changer )

Photo Facebook

Vous commentez à l'aide de votre compte Facebook. Déconnexion / Changer )

Photo Google+

Vous commentez à l'aide de votre compte Google+. Déconnexion / Changer )

Connexion à %s