Artwork

Conteúdo fornecido por Winfried Adalbert Etzel - DAMA Norway. Todo o conteúdo do podcast, incluindo episódios, gráficos e descrições de podcast, é carregado e fornecido diretamente por Winfried Adalbert Etzel - DAMA Norway ou por seu parceiro de plataforma de podcast. Se você acredita que alguém está usando seu trabalho protegido por direitos autorais sem sua permissão, siga o processo descrito aqui https://pt.player.fm/legal.
Player FM - Aplicativo de podcast
Fique off-line com o app Player FM !

3#4 - Anders Dræge - AI-driven Process Automation (Nor)

37:55
 
Compartilhar
 

Manage episode 379346163 series 2940030
Conteúdo fornecido por Winfried Adalbert Etzel - DAMA Norway. Todo o conteúdo do podcast, incluindo episódios, gráficos e descrições de podcast, é carregado e fornecido diretamente por Winfried Adalbert Etzel - DAMA Norway ou por seu parceiro de plataforma de podcast. Se você acredita que alguém está usando seu trabalho protegido por direitos autorais sem sua permissão, siga o processo descrito aqui https://pt.player.fm/legal.

«I think that having a very good framework, where you can put all ML and AI in, makes it much easier, much more clear. (Jeg tror det å ha et veldig bra rammeverk, der du kan putte all ML og AI inn i, det gjør at du får det mye lettere, mye mer oversiktlig.)»
Frende Forsikring, a Norwegian Insurance Company has build up a team of 6 people that work with Machine Learning (ML) and Artificial Intelligence in the company. Their goal is to ensure the companies growth through automation. Anders Dræge is the Head of the Machine Learning and Artificial Intelligence team at Frende Forsikring and he has always had an interest for data and automation.
Anders is not just an award winning Data Scientist, one of the Nordic 100 in 2023, but also a person that is happy to share his knowledge.

The goal for Automation

  • Automation is a target that can be measured against
  • You can measure both, time saving as well as saved cost
  • High-risk items are a good use-case to show the effect of ML: Its not necessarily about replacing work tasks, but to ensure that human focus in on the items that are of highest risk and value
  • Automation is a way of scaling and growing your business, without increasing resources.
  • The need for automation becomes more clear, and to avoid over-allocation of resources, the need is evident in the business.
  • Your goals fro AI and automation have to be aligned with your organizations business goals

The composition of the team

  • The Machine Learning team is 6 people string, consisted of
    • 2 ML engineers
    • 2 are 50% actuary (domain knowledge connection)
    • 1 data engineer that prepares data
    • 1 MLOps developer with interest in ML to build connections with IT department
  • Close collaboration with RPA (Robotic Process Automation) team and other departments.

The process

  • The trinity of data in ML is paramount for quality results:
    • 1 set to train
    • 1 set to validate
    • 1 set to test
  • There are possibilities to automate testing procedures
  • Monitoring can and should be automated

The technological framework

  • Find a framework that can control your processes, detect deviations and monitor effectively.
  • Implementation and setting things in production is much more efficient with a proper framework
  • Find a standard way of operating, will also have a positive effect on on-boarding new people

Key factors for success

  • «One factor that was decisive for a very good collaboration across teams and departments is that we are very close. (En faktor som var avgjørende for et veldig godt samarbeid på tvers av team og avdelingene, er det at vi sitter veldig nært.)»
  • Physical co-location is a success factor
  • A lot of key competency is in-house
  • Clear and transparent message on automation
  • A culture that is actively striving for automation, finding ways to improve
  • Culture is really important: People have to be receptive to the ideas of automation
  • Find the right time to talk about automation - ideally before the need arises
  • Human in the loop
  • Monitoring of process output by humans is important for most ion the processes. This is about evaluating output with expectations from human experience
  • Human evaluation becomes input for re-training of the model

The use cases

  1. Automatic email distribution
  2. Processing of physical mail
  3. Monitoring of laws

For the work Frende Forsikring has done with Natural Language Processing (NLP) for email distribution, the team won the Dataiku Frontrunner Award 2023.
https://www.frende.no/aktuelt/frende-vant-internasjonal-ai-konkurranse/

  continue reading

69 episódios

Artwork
iconCompartilhar
 
Manage episode 379346163 series 2940030
Conteúdo fornecido por Winfried Adalbert Etzel - DAMA Norway. Todo o conteúdo do podcast, incluindo episódios, gráficos e descrições de podcast, é carregado e fornecido diretamente por Winfried Adalbert Etzel - DAMA Norway ou por seu parceiro de plataforma de podcast. Se você acredita que alguém está usando seu trabalho protegido por direitos autorais sem sua permissão, siga o processo descrito aqui https://pt.player.fm/legal.

«I think that having a very good framework, where you can put all ML and AI in, makes it much easier, much more clear. (Jeg tror det å ha et veldig bra rammeverk, der du kan putte all ML og AI inn i, det gjør at du får det mye lettere, mye mer oversiktlig.)»
Frende Forsikring, a Norwegian Insurance Company has build up a team of 6 people that work with Machine Learning (ML) and Artificial Intelligence in the company. Their goal is to ensure the companies growth through automation. Anders Dræge is the Head of the Machine Learning and Artificial Intelligence team at Frende Forsikring and he has always had an interest for data and automation.
Anders is not just an award winning Data Scientist, one of the Nordic 100 in 2023, but also a person that is happy to share his knowledge.

The goal for Automation

  • Automation is a target that can be measured against
  • You can measure both, time saving as well as saved cost
  • High-risk items are a good use-case to show the effect of ML: Its not necessarily about replacing work tasks, but to ensure that human focus in on the items that are of highest risk and value
  • Automation is a way of scaling and growing your business, without increasing resources.
  • The need for automation becomes more clear, and to avoid over-allocation of resources, the need is evident in the business.
  • Your goals fro AI and automation have to be aligned with your organizations business goals

The composition of the team

  • The Machine Learning team is 6 people string, consisted of
    • 2 ML engineers
    • 2 are 50% actuary (domain knowledge connection)
    • 1 data engineer that prepares data
    • 1 MLOps developer with interest in ML to build connections with IT department
  • Close collaboration with RPA (Robotic Process Automation) team and other departments.

The process

  • The trinity of data in ML is paramount for quality results:
    • 1 set to train
    • 1 set to validate
    • 1 set to test
  • There are possibilities to automate testing procedures
  • Monitoring can and should be automated

The technological framework

  • Find a framework that can control your processes, detect deviations and monitor effectively.
  • Implementation and setting things in production is much more efficient with a proper framework
  • Find a standard way of operating, will also have a positive effect on on-boarding new people

Key factors for success

  • «One factor that was decisive for a very good collaboration across teams and departments is that we are very close. (En faktor som var avgjørende for et veldig godt samarbeid på tvers av team og avdelingene, er det at vi sitter veldig nært.)»
  • Physical co-location is a success factor
  • A lot of key competency is in-house
  • Clear and transparent message on automation
  • A culture that is actively striving for automation, finding ways to improve
  • Culture is really important: People have to be receptive to the ideas of automation
  • Find the right time to talk about automation - ideally before the need arises
  • Human in the loop
  • Monitoring of process output by humans is important for most ion the processes. This is about evaluating output with expectations from human experience
  • Human evaluation becomes input for re-training of the model

The use cases

  1. Automatic email distribution
  2. Processing of physical mail
  3. Monitoring of laws

For the work Frende Forsikring has done with Natural Language Processing (NLP) for email distribution, the team won the Dataiku Frontrunner Award 2023.
https://www.frende.no/aktuelt/frende-vant-internasjonal-ai-konkurranse/

  continue reading

69 episódios

Todos os episódios

×
 
Loading …

Bem vindo ao Player FM!

O Player FM procura na web por podcasts de alta qualidade para você curtir agora mesmo. É o melhor app de podcast e funciona no Android, iPhone e web. Inscreva-se para sincronizar as assinaturas entre os dispositivos.

 

Guia rápido de referências