By Yashoda Singh Solution spesialist - Data og AI, Microsoft Norge | Leder for programkomiteen for Make Data Smart
09:05 - 09:35
AI Revolution Again, what’s different this time?
By Magnus Revang Chief Product Officer at Openstream.ai
09:35 - 10:00
AI Act - What happens when technology surpasses laws
By Tore Tennoe Director of the Norwegian Board of Technology
10:30 - 11:00
Helping save lives with AI
By Alexander Dahl Vice President Data and Analytics, Laerdal Medical AS
Helping save one million more lives by 2030 - that the mission of Laerdal Medical. A world leader in medical simulation, education and training, Laerdal Medical is working with our partners to develop solutions using advances in digital, data and AI to reach this goal and beyond.
From helping save newborns in Tanzania and Nepal, to assessing training efficiency and outcomes in USA, and self-directed CPR training in UK - this is a short and inspirational talk about how data, analytics and AI plays a key role in helping save lives across the globe.
#3 Smart
10:30 - 11:00
Simplifying and speeding up the analytical life-cycle using a unified analytical and transactional database
This presentation explains how to simplify the data- landscape, model, and processing using a unified analytical and transactional database to store all types of data within the same database and ingest and process data in real-time.
Moving the analytical engine into the database and removing the data movement speeds up all the analytical life cycle steps and simplifies the users’ workprocess. SAS’s experience implementing SAS Viya with Singlestore with customers is the foundation of this presentation. It covers designprinciples and architecture with concrete examples of the technology used.
#2 Data
10:30 - 11:00
Revolutionizing Dagens Næringsliv with Language Models: A Journey of both Successes and Failures
By Martin Kermit Head of Data Science, DN Media Group
The realm of natural language processing has seen an explosion of practical applications in recent years. It's no longer just
about simple spell checks or social media monitoring - we now have advanced chatbots like GPT, capable of generating text
and engaging in human-like conversations.
In the media industry, technologies related to language modelling have acted as catalysts for significant evolution. These
advancements present immense opportunities to enhance product quality, streamline news production, and elevate user
experiences. However, while this technology holds tremendous promise, it's not without its challenges. Misaligned
expectations or overambitious aims can lead to disappointment and frustration. As with any innovative technology, it's vital to
understand its capabilities and limitations to truly harness its potential.
Five years ago, the DN media group embarked on our journey with natural language processing. It was a path of learning,
paved with initial missteps that prompted us to rebuild from scratch, developing necessary infrastructure and fostering
expertise within our organization one step at a time. Today, we've successfully implemented automated systems that, assist
reporters throughout the publication process, classify news topics as well as using the generative capabilities of GPT-models
in a controlled environment. This talk will provide a detailed look into our transformative journey.
I dag skal «alle» bli datadrevet, men hvordan skal man stykke opp den elefanten? Det handler litt om teknologi, men mest om menneskene og prosessene våre. Hvordan bygge organisasjon og kompetanse som skaper en datadrevet kultur? Her gir jeg et innblikk i vår datadrevet reise i Nortura samt hvordan vi klarer å holde øye på målet underveis (spoler: vi er ikke ferdig). I gode datadrevet ånd så deler jeg flere matnyttig råd i form av feil og suksesshistorier, og skal unngå å snakke i store floskler.
Data Mesh, and distributed data management is a hot topic these days, but for many organizations the gap between today’s centralized setup and the fully distributed Data Mesh is large and hard to approach. At reMarkable we have tried to balance the theoretical upsides of distributing the data management with the organizational realities, into what we have called the tailored Data Mesh. Arguably we have (after some trial and error) found a well working balance that is fit for purpose now, and well rigged for future adjustments towards a more distributed setup as the data maturity increases.
In 2022, triggered by increasing growth and a radical change in business models, reMarkable decided to set up a new Data & Insight function. When designing the target state for this function, we looked towards the Data Mesh principles for inspiration, but also consider the organizational realities and key drivers arguing for a more traditional and centralized approach – at least as a starting point. In this presentation we will start by sharing how we approached the process of designing the operating model, what we decided on and why, and then the experience (positives and negatives) one year after.
The presentation will be in Norwegian.
#2 Data
11:00 - 11:30
Developing production-ready large language models. Is it just engineering?
Large language models (LLMs) are enabling Data Scientists to build applications that they previously could not. Many businesses are experimenting with prototype applications powered by LLMs. Moving LLM applications from PoC into production can be challenging because the required processes, architectures and tools are still in their infancy. In this talk we will explore some of the challenges we have seen when building an Enterprise Level LLM solution as well as the intersection between Data Science and Software Engineering for this kind of projects.
#3 Smart
11:30 - 11:45
Break
11:45 - 12:15
ESG Data Hub: A story about converging fronteers between sustainability and technology
The talk will describe DNB's journey on how the ESG Data Hub was created and what is being used for, and future plans. The ESG Data Hub is a modern, cloud-based ""Hub"" that stores and provides access to a wide range of data and analytics or insight services pertaining ESG (Environmental, Social, and Governance) aspects. The Hub connects to external data sources, would they be public sources such as the National Registry of Companies or Statistics Norway, or third-party data vendors, and integrates this external ESG data to DNB's internal systems, such as customers, loans or suppliers.
The 'Hub' is one of DNB's success stories on their journey to adopting the Data Mesh approach across the entire group. The Hub has as mandate to provide in an efficient, compliant and ethical way, data and analytics across the entire group, for solving different use cases such as: External reporting (such as annual reports), input for risk models (e.g., risk models would have to account for climate risk in the near future), new and better products and services for our customers (e.g., offering green loans, insights for being more sustainable, etc.), as well as understanding better the ESG risks and opportunities at Customer, Transaction and Portfolio levels.
#2 Data
11:45 - 12:15
Hvordan digital transformasjon utfordret en toppledergruppe. Erfaring med NAVs reise 2010-2015
Vi i toppledergruppen var kanskje de mest inkompetente da vi skulle modernisere NAV til å bli en digital organisasjon. Vi var for umodne til å skjønne hvor umodne vi var." Det er vanskelig å ta de riktige beslutningene når man ikke forstår det man driver med. Økt egen kompetanse er et nøkkelord, men også organisering slik at man får riktige råd.
I denne presentasjonen demonstrerer Hans Martin hvordan BAMA har tatt i bruk matematisk optimering for
transportplanlegging, og hvordan datadrevne beslutninger har resultert i mer effektiv vareflyt, reduserte kostnader, og ferskere
frukt og grønt.
Du vil få svar på:
- Hva er matematisk optimering?
- Hva brukes optimering til i BAMA?
- Hvordan har BAMA tatt optimeringsmodeller fra POC til produksjon?
#3 Smart
12:15 - 12:45
How to Build a Data Science Function and Deliver Value to the Business from Day 1
By Jari Kunnas Head of the Data Science Centre of Excellence, Vår Energi, Jonathan Kirk Senior Data Scientist, Capgemini
Ever wondered how to get value from data science in your organisation? Seen other companies doing amazing things and
wanted your own piece of the action?
Let us take you through our journey from the very start and show you how we navigated the challenges and ensured that
delivering value to the business was at the heart of everything we did!
And for those of you with an already established data science function, there’s our key takeaways to ensure that at every
stage of the journey, you get the most out of Data Science in your organisation.
#1 Make
12:15 - 12:45
When AI is used to build AI - what could go wrong?
By Jon Jahren Director, Azure Cloud & AI, Microsoft
What if you could converse with your data and ask it to reveal its secrets? What if you could use natural language and AI to create data and AI solutions that can help you solve complex problems and make better decisions? What could go wrong? The next generation analytics platforms will introduce AI copilots, using advanced generative AI to help you discover and share insights and build data & AI solutions faster. In this session we discuss how your lives will change to becoming AI prompt engineers, and the benefits (and potential pitfalls) that this will bring.
#3 Smart
12:15 - 12:45
SINTEF OceanLab: Enabling sustainable innovation from marine data streams through data virtualization
SINTEF OceanLab has complex and varied data sources on several locations. With the use of Data Virtualization technology they will succeed in making a dataplattfrom that enables a smarter and environmentally optimized data platform for data sharing both for internal and external resources and systems.
#2 Data
12:45 - 13:45
Lunch
13:45 - 14:15
Fra strømlinjer til datastrømmer - data og analyse for samfunnskritisk infrastruktur
I jobben med å drive det norske kraftsystemet og sørge for at du har strøm i kontakten til enhver tid samler Statnett inn store mengder data. Men hvordan få en organisasjon til å ta disse dataene i bruk når kvaliteten i beste fall er udokumentert, vi mangler kompetanse og systemene våre må virke hele tiden? Dette er historien om hva vi har fått til, hvor vi har feilet og hva dette har lært oss om mennesker og data.
#2 Data
13:45 - 14:15
Data driven dynamic pricing for short-term vacation rentals
By Alex Lenkoski Chief Research Scientist, The Norwegian Computing Center
We discuss an eight-year long collaboration between Norsk Regnesentral (NR) and San Francisco-based startup Wheelhouse. Wheelhouse allows owners of short-term vacation rental properties (Airbnb, vrbo, etc.) to dynamically price their units, automatically adapting to changing demand environments in order to maximize revenue. Sitting behind this interface is an enormous and constantly expanding dataset of observed market behavior, along with a pricing framework developed by NR. This framework is a blend of "old" and "new" ideas, combining 1970's style economic models of demand in markets for heterogeneous goods, survival time methods borrowed from biostatistics and machine learning techniques. We will highlight some of the successes of our methodology and also review how these methods adapted to the instabilities introduced throughout the COVID era as well as discuss the potential of this platform moving forward.
If you have used Ruter's offer, you have probably used an AI algorithm. He will share how Ruter uses artificial intelligence to
develop its services, follow customer needs and drive innovation in the transport sector.
#3 Smart
14:15 - 14:45
Mapping a Path to Ocean Sustainability: The Role of Big Geospatial Data and Sensors
HUB Ocean is an independent, non-profit, non-compete foundation. We are a partner community developing a ground-breaking data platform, applications and tools to pilot new approaches to ocean governance. Our mission is “To change the fate of the ocean by unleashing the power of data, technology and collaboration”. We are developing the Ocean Data Platform in close collaboration with partners from science, governments and industries. In addition, we have important technology partners like Microsoft (with the AI for Good lab and more) as well as Cognite.
By leveraging big geospatial data, which includes satellite data, aerial surveys, unmanned surface vessels, autonomous underwater vehicles, seabed-landers, profiling floats and much more, we can gain unprecedented insights into the state of our oceans. These data sources provide detailed information on ocean currents, water quality, biodiversity hotspots, and the impacts of human activities. Additionally, sensors deployed in various marine environments enable real-time monitoring, helping us understand dynamic changes and respond effectively to emerging challenges.
During this presentation, we will delve into the innovative applications of big geospatial data and sensor technologies in ocean sustainability. We will showcase case studies where these tools have been employed to support marine conservation, optimize resource management, and mitigate risks associated with climate change. Furthermore, we will explore the potential of artificial intelligence and machine learning techniques to extract actionable insights from vast ocean datasets, facilitating evidence-based decision-making.
Join us to discover how the integration of big geospatial data and sensors can pave the way for a more sustainable and resilient future for our oceans. Together, we can unlock the power of data-driven solutions and foster collaboration among stakeholders to protect and preserve our precious marine ecosystems.
#2 Data
14:15 - 14:45
Navigating the path to Trustworthy AI
By Josefin Rosén Trustworthy AI specialist, SAS Institute
Artificial intelligence (AI) has become increasingly essential for critical decision making processes in both corporate and
societal contexts. Additionally the recent introduction of generative AI, such as ChatGPT, has put the power of AI in the hands
of users and sparked a surge of interest in its potential to transform various aspects of our lives and open up exciting new
opportunities for businesses.
However, the rapid rise of AI has also underscored the need for responsible and ethical development and deployment of
these technologies. Without proper governance, AI can result in unintended ethical challenges that can quickly spiral out of
control.
To address these concerns, the EU Commission has proposed the AI Act as a significant step towards promoting responsible
and trustworthy AI on a larger scale.
In this presentation, we will provide recommendations for promoting responsible,
sustainable AI-driven innovation. Join us to prepare for the future of AI and ensure that it is deployed in a manner that benefits
society and fosters fairness and equality.
#3 Smart
14:15 - 14:45
Better Sustainability with IoT solutions for Data-driven asset management at Bane NOR
By May-Britt Støen Project Manager, Bane NOR, Håkon Lenschow Managing Consultant & Team Lead, Capgemini, Martin Lam Head of Technology, Principal Solutions Architect, Capgemini Insights and Data
Sustainability is a key word for a better future. Bane NOR has a clear vision for how they can contribute to a more sustainable
society in general and a more sustainable transport sector. One of their strategies for sustainability 2021-2025 is " Mer på
skinner setter mindre spor". In this presentation, we will talk about how Bane NOR has built solutions based on IoT-data to
move from planned asset management to data driven asset management in order to be more sustainable. This is about how
Bane NOR get sensor data from trains, buildings and other facilities, contextualize them and use the insights from those data
to make smarter decisions on how, when and where they should invest their energy, money and resources.
All the solutions are built on Azure. We will share our solution architectures, best practices and learned lessons to the
audience. Examples for some technology components that have been used are Azure Synapse, Spark, Azure Event Hub,
Azure Data Explorer, Azure data lake, Azure SQLDB, Grafana and Power BI
#1 Make
14:45 - 15:00
Break
15:00 - 15:30
DevOps revolutionised software development. It's time to revolutionise data.
By Alexandra Diem Head of Cloud Analytics and MLOps, Gjensidige Forsikring
What does a team of developers do in a BI and Analytics environment? The use of data analysis within a modern technology stack is at the core a successful insurance business. Gjensidige, Norway's largest insurance company, has recently begun it's journey of moving all data operations from an on-prem data warehouse solution into the cloud, and with over 20 different analyst teams in the mix, that is no mean feat. Everything we do is data-driven with the goal to provide the best possible customer experience: from pricing to marketing and sales, as well as direct customer service, and claims handling. To succeed in this journey, we utilise the power of software best practices and DevOps culture: Product thinking, end-to-end responsibility, cross-team collaboration, and early problem solving. In this talk I will show how employing developers to both grow and root DevOps culture and software best practices within our analyst teams has been a crucial ingredient to our success in publishing and maintaining high quality data products into our data platform.
#2 Data
15:00 - 15:30
Becoming twoday: Using AI to make large organisational changes a bit less painful
By Per Anders Waaler Analytics Architect and Advisor, Twoday, Joachim Aae Analytics Architect and Advisor, Twoday
twoday startet som en sammenfatning av et tyvetalls individuelle selskaper, og har et
mål om å bli et enhetlig nordisk teknologiselskap. En del av dette innebærer å migrere
2500 ansatte fra Google som samarbeidsplattform til Microsoft.
For å gjøre migreringen så sømløs som mulig, har twoday benyttet sin egenutviklede
løsning for bedriftsspesifikke AI-assistenter basert på Azure OpenAI og GPT. Ved å gi
modellen tilgang på intern dokumentasjon, kunne den svare på de vanligste
problemstillingene som ansatte hadde, slik at de ikke måtte bruke tid på å lese gjennom
mye dokumentasjon. I tillegg har twoday brukt AI Agenten i flere interne
forretningsprosesser – både for å bli mer datadrevet, og for å utnytte
produktivitetsgevinster av AI som ikke har vært mulig før den siste utviklingen av
språkmodeller som GPT.
Per Anders Waaler og Joachim Aae vil presentere bruken av en AI-basert assistent som
ledd i en større organisatorisk endring. Hva fungerte godt, og hvilke fallgruver støtte de
på som andre kan lære av?
#1 Make
15:00 - 15:30
Integrating Machine Learning, Computer Vision, and Photogrammetry for Advanced Drone Image Processing + Biodrone
By Paul Tenfjord + Biodrone Senior Data Scientist, Sopra Steria + Leading expert in custom drone missions, Biodrone
Biodrone is revolutionizing forestry management with its innovative drone technology. By integrating AI, they've created a
portal for analyzing forests and farmland using drones that leads to sustainable precision forestry. The traditional, timeconsuming method of on-foot inspection is replaced with a detailed aerial overview, offering accurate insights such as the
number and average height per tree species. This advanced approach provides much-needed insights in a quick, costeffective, and eco-friendly manner.
This accessible web based portal, developed by Sopra Steria, requires no prior knowledge of GIS, photogrammetry, or AI,
making it a game-changer for everyone in the industry. From forest owners to contractors and administrators, users can
simply upload drone-captured images or videos for processing and analysis. By utilizing state-of-the-art AI technology, the
system recognizes different forest and plant species, and creates a digital twin of your forest, from images taken at up to
200m height.
#3 Smart
15:30 - 16:00
Hvordan du klarer å skape verdier med AI?
By Magnus Glader Global Digital Advisor, Columbus Data&Analytics.
Det er lett for AI-prosjekter å ende opp som tekniske piloter og aldri bli vellykket implementert i virksomheten. Basert på våre erfaringer med vellykkede AI-prosjekter viser vi hvordan du kan realisere verdien av dine AI-investeringer.
#1 Make
15:30 - 16:00
En ML-reise mot automatiske ekspansjonsanbefalinger
SpareBank 1 Forsikring achieves flexibility to quickly deliver data to support analytical user needs and implement new use cases with real-time data
By Håkon Tveita Director/CIO, SpareBank 1 Forsikring
Performance was one of the key challenges that SpareBank 1 Forsikring faced in the past. The organization is dealing with sensitive client data, and needs to be compliant with regulations (i.e. GDPR and AML). Timely delivery of complex data sets was indispensable, and the company was looking to replace a legacy solution in order to enable rapid implementation of new data. With the help of Denodo, SpareBank 1 Forsikring was able to mask data in accordance with GDPR and work with key users in establishing clear “rules of engagement” and common terminology. The solution enabled the company to prepare for future cloud migration without breaking “data-contract” with the users. The benefits of the “Information platform” became clear when developing new solutions such as the anti-money laundering solution.
#2 Data
16:00 - 16:20
Break
16:20 - 16:45
Panel Discussion: Revolutionising organisational Practices with data & AI: Overcoming challenges and Maximizing Potential
By , , ,
Revolutionising organisational Practices with data & AI: Overcoming challenges and Maximizing Potential.
#0 Plenum
16:45 - 17:00
Closing Remarks
By Yashoda Singh Solution spesialist - Data og AI, Microsoft Norge | Leder for programkomiteen for Make Data Smart