Skip to main content

Schlagwort: Databooster

Expert Group Meeting – Data Ethics

The Expert Group will be meeting at the Digital Society Initiative to discuss the outlining project goals of their project AI Law Supporter (AILAS).

Expert Group Meeting – Machine Learning Clinic: Behind the Hype GenAI – Real Applications and Benefits for Industry?

Dear ML Clinic experts, 

I am happy to inform you that we are organizing our next ML Clinic workshop which will be hosted by Swisscom. We will have presentations from Swisscom and CSEM followed by an interactive panel discussion to develop further on this exciting topic:

Behind the Hype GenAI” – Real applications and benefits for industry?

Please find attached the flyer with all the details about the event.

Agenda

16:00 – 17:00 – Impulse Presentations

From Idea to Impact – Identifying and Developing Scalable AI Use Cases
Mattmann Carole – Senior Data & AI Consultant, Swisscom AG
An approach to develop AI-driven solutions, guided by a structured framework that supports scalable development and provides a clear methodology for identifying and defining impactful use cases. Through real-world examples across various industries, this presentation will demonstrate how our framework helps us navigate the complexities of AI deployment, including model scalability, data integration, and operationalization, to achieve meaningful business results.

Discover the Swiss AI Platform
Levy Sarah – Stream Lead AI Strategy & Offerings, Swisscom AG
Switzerland’s AI landscape requires balancing innovation with strict data governance. The Swiss AI Platform, a Swiss Army knife for AI, meets this need by offering GPU-as-a-Service, an AI Workhub for collaboration, and a GenAI Studio for generative AI, all underpinned by Swiss data governance and Nvidia’s scalable technology. Discover how this all-in-one platform enables secure, high-performance AI solutions, driving new possibilities for Swiss innovation.

The impact of Generative AI in Industrial Applications
Iason Kastanis – Group leader Predictive Analytics, CSEM SA
This talk will describe CSEM projects in the area of multimodal industrial data with examples from real world applications. Many years of experience have highlighted the idiosyncrasies of the digitalisation process taking place on shop floors across the industry. The rise of GPT-based large language models presents an opportunity to use Generative AI to address the issues faced in the development of AI-driven analytics for the industry.

Defining the boundaries of AI From fundamental limitations to resource constraints
Nadim Maamari – Group Leader Edge AI and Vision System, CSEM
Defining the boundaries of AI involves recognizing its intrinsic limitations, such as lack of true understanding and dependence on data, while balancing its immense computational power needs, which pose sustainability challenges. As AI grows more powerful, ethical considerations and regulatory measures become essential to ensure responsible and energy-efficient advancements, guiding its impact on society and the environment

17:00 – 17:45 Panel Discussion and Q&A

Challenges and opportunities for the Swiss industry
Mattmann Carole, Levy Sarah, Iason Kastanis, Nadim Maamari, and Philipp Schmid (Head Manufacturing Industries, Swisscom AG)
Moderator: Francesco Crivelli, Head of Research & Business Development for Industry 4.0 and ML, CSEM

17:45 – 19:00 Apéro Riche and Networking

Workshop on the Initiative PIVOT: Political Insights & Views Over Time

PIVOT: Political Insights & Views Over Time is a new research initiative that aims to track the evolution of political opinions and public stances over time. 

In this context, we are seeking collaborators who bring diverse perspectives from both academic and industry backgrounds and are interested in participating in a half-day workshop. This event will serve as a collaborative platform to better understand the needs and expectations of all involved parties, and align our project goals with these.

What is PIVOT?

The initiative PIVOT is designed to analyze and map out the shifts in the opinions of politicians and public figures, offering insights into why and how their positions on key issues evolve over time. With the power of modern AI methodologies, most notably Large Language Models, and access to comprehensive datasets (SwissdoxCuria Vista etc.), we are equipped to track and interpret long-term opinion shifts across different domains and formats, including PDFs, speeches, and public statements.

Why This Project Matters:

  1. Fighting Fake News: In today’s fast-paced digital world, misinformation can spread faster than facts. Political statements are often taken out of context, which can fuel fake news and disinformation. By analyzing long-term trends in political stances, PIVOT offers a more accurate, contextualized view of politicians‘ opinions over time, helping both the public and media outlets verify the authenticity of statements and counteract misinformation.
  2. Democracy at Stake?: With democracy being challenged globally, understanding political accountability is more crucial than ever. The project aims to create transparency around political shifts, allowing voters, policymakers, and industry leaders to track how politicians‘ views evolve over time, beyond the typical 4-5 year election cycles.
  3. Long-Term Political Analysis: Politicians often focus on short-term outcomes, usually up to the next election cycle. PIVOT, however, takes a broader view. By using longitudinal data analysis, we aim to uncover patterns of opinion changes over years or even decades and explore the key events or influences that drive these changes.
  4. Public Relations & Media: For media companies or public relations teams, PIVOT’s analysis offers valuable insights into how politicians’ public stances align with or deviate from historical positions. This can be leveraged to craft informed messaging, fact-check statements, and anticipate future shifts.
  5. Risk Management & Consulting: Political shifts often influence business strategies, regulatory frameworks,and market stability. Partnering on this project could offer data-driven insights for businesses seeking to navigate evolving political landscapes, particularly for industries like energy, finance, or healthcare.

Meet Fellow Experts

By joining PIVOT, you will have the opportunity to collaborate with leading experts in political sciencedata analytics, and AI. Together, we aim to generate actionable insights that can benefit not only academia but also industries that rely on understanding political riskpublic sentiment, and long-term policy trends.

Next steps

To bring PIVOT to life, we will conduct the first half-day workshop that will take place on November 25th 2024 in Bern.

We aim to gather a diverse group of potential partners from academia, industry, media, and other relevant sectors to discuss the project’s concept in detail, exchange ideas, and explore how different organizations can contribute their expertise and resources. The workshop will help us refine the objectives and identify key areas of research where collaboration is essential. Additionally, it will provide an opportunity for participants to network and connect with others working in areas such as political data analysismedia fact-checking, and longitudinal political insights.

During the workshop we will also elaborate the next steps for continuing and supporting PIVOT as a long-term project.

What we need from you

If you are interested in the project and would like to take part in the mentioned workshop, please fill in and send us this feedback and contact Veton Matoshi (msv3@bfh.ch). We would like to know about your competencies and experiences relevant for the potential project.

We value every expression of interest and potential contribution to this project. To ensure a productive and focused workshop, we’ll carefully review all submissions and send official invitations to participants whose expertise and experience align closely with the project’s current needs and direction. We appreciate your understanding as we strive to create a balanced and effective collaborative environment for this initial workshop.

If you have any questions or require further information, please don’t hesitate to contact us at any time.

Workshop – From Crowdsourced to Reliable Data: AI Quality Evaluation of OpenStreetMap

OpenStreetMap (OSM) is a valuable crowdsourced alternative to commercial and governmental geospatial data providers, with proven applications e.g. in the environment, tourism and emergency services. However, the lack of quality assessment tools hinders its wider adoption. This workshop aims to address the need for reliable and standardized data quality evaluation in OSM, with a focus on AI-based methods.

Current research prototypes such as ‚OSM Data Completeness‘ and ‚ohsome quality analyst‘ offer intrinsic solutions for data quality evaluation in specific regions based on user requirements. But there’s potential for improvement in usability and integration of extrinsic approaches.

The workshop will explore how these tools can be developed for production use and identify areas for further research. By discussing data quality evaluation solutions, the workshop aims to enable the establishment of an applied research and development agenda to better evaluate the suitability of OSM data.

The workshop will be held in German and English.

Registration and contact: Mail to stefan.keller@ost.ch

Tage der Technik – Generative AI Use Cases im Smart Service Umfeld

Die Fachveranstaltung bietet einen exklusiven Einblick in einen preisgekrönten Generative AI Use Case, der Echtzeitdaten und Knowledge Management im IoT-Bereich verbindet. Dabei werden Fehlermeldungen und Anlagedaten mit strukturierten und unstrukturierten Firmendaten verknüpft, um so dem Benutzer einen smarten Self-Service zu bieten. Zusätzlich werden weitere innovative Generative AI Use Cases und Anwendungen für Unternehmen vorgestellt.

Referent: Chris Bochsler, Managing Partner Cando Image GmbH

Jürg Meierhofer und Jochen Wulf führen durch die Veranstaltung.

Im Anschluss an die Vorträge offerieren wir den Teilnehmenden ein Sandwich und Getränk. Dazu können Sie sich mit den anderen Gästen und den Referenten austauschen.

Eintritt
Der Eintritt ist frei
Die Plätze sind beschränkt, bitte melden Sie sich für die Veranstaltung an:

Expert Group Meeting – Natural Language Processing in Action: Contextualization

The next expert meeting of the NLP group will be held on 1. October 2024 at the usual location – ZHAW on Lagerstrasse – followed by an Apero at Nüü. This session will focus on the topic of „Contextualization,“ featuring the following speakers:

  • Hans Ramsl from Weights and Biases will discuss “Beyond Vibe Checks: Establishing Systematic Evaluation Frameworks for Semantic Contextualization in LLMs”.
  • Tobias Schimanski, PhD candidate at the University of Zurich and affiliated researcher at the University of Oxford, is going to talk about “Challenges and Pathways for Retrieval Augmented Generation” by looking into two of his recent papers.
  • Dietrich Trautmann, AI Engineer and Creator of Assistant Shell, will present on „Context-Driven Development: Exploring the Potential of LLMs in Software Engineering“. Dietrich will explore how leveraging Large Language Models (LLMs) can revolutionize software engineering practices, focusing on context-aware development workflows, enhanced code generation, and intelligent error handling.

Join us on October 1st at 17:30 at Lagerstrasse 45, 8004 Zurich, in room ZL O6.16 (6th floor). The presentations will run from 17:30 to 19:00, followed by an Apero downstairs at 19:00.

Online participation is also possible.

Please use the following form to confirm your attendance by September 24

We will send you a calendar invite which also includes online participation details.

maintenance Schweiz 2024

Die Maintenance Schweiz ist die Leitmesse für industrielle Instandhaltung

Vom 28. – 29. August 2024 findet die 16. Ausgabe der maintenance Schweiz in Zürich statt. Die Schweizer Leitmesse für industrielle Instandhaltung ist ein fixer Bestandteil für Alle, die sich mit Anlagenwartung, Ersatzteilen, Smart Maintenance, Arbeitssicherheit und technischen Lösungen auseinandersetzen. Die Messe präsentiert Lösungen und Trends für die gesamte Schweizer Industrie.

Die data innovation alliance und der Databooster werden am smart data forum präsent sein und in mehreren Talks den Nutzen und Anwendung von smart data vorstellen. Die Registrierung ist kostenlos!

Digital Health Lab Day 2024

Dear Digital Health Enthusiasts

The next Digital Health Lab Day will take place on Tuesday, 3rd September 2024 at the ZHAW School of Engineering in Winterthur. Join participants from healthcare, IT, management, health experts, startups and research, learn and discuss about recent developments in digital health.

Find more information and the registration on the official website.

Databooster – Finale

Join us for the Databooster Finale as we celebrate the achievements of the four-year Innovation Booster program, which concludes at the end of this year.

Save this date as well! Details will be updated before the date.

Webinar – The role of documents in RAG Applications

Join us for an insightful webinar where we explore how to transform a company’s documents into structured data, paving the way for advanced Generative AI applications.

As member of Data Innovation Alliance in this educational webinar we will guide you through the process of converting diverse document formats into usable data. You’ll learn about advanced data extraction techniques, ensuring data accuracy and traceability, and see real-world applications in Retrieval Augmented Generation (RAG). Ideal for professionals aiming to leverage AI for augmented reading, chat, analytics, and automation.

Key Takeaways:

  • Understanding document complexity and data extraction
  • Techniques for ensuring data accuracy
  • Practical RAG application insights

Agenda:

  • 5 min  –  Welcome and Introduction 
  • 15 min – Key Challenges and learnings of extracting data from unstructured sources 
  • 15 min – Practical use cases in augmented reading, RAG 
  • 10 min – Q&A

The webinar will be recorded for on-demand watching.

Abonnieren Sie unseren Newsletter: