AI@WORK


AI@WORK: A Workshop at the Informatik Festival 2024

AI@WORK has been held on 24.09.2024 at the Informatik Festival 2024 / Wiesbaden. Thanks to everyone for contributing and participating!

The workshop brings together AI researchers and practitioners interested in innovative AI applications for working environments across all sectors. It has included 11 presentations, and has offered the opportunity for peer-reviewed publication.

Program

Here is the program (last update: Sep 17 2024).

Time Topic Speaker
09:30 - 09:40 opening Adrian Ulges
Session "AI for Knowledge Work" Jörn Hees (chair)
09:40 - 10:00 Expanding Knowledge Graphs Through Text: Leveraging Large Language Models for Semi-Inductive Link Prediction Felix Hamann
10:00 - 10:20 Using Large Language Models to Generate Authentic Multi-agent Knowledge Work Datasets Desiree Heim
10:20 - 10:40 Towards Context-aware Recommender Systems for Supporting Knowledge Workers in Personal and Corporate Information Space Mahta Bakhshizadeh
10:40 - 11:00 Application of Graph Neuronal Networks to Fraud Classification Problems in Financial and Insurance Domain Jona Becher und Andreas Schäfer
- coffee break -
Session "AI for Software Development" Christian Jilek (chair)
11:30 - 11:50 Evolving Software Development: The Assisting and Transforming Role of Generative AI Constantin von Brackel-Schmidt
11:50 - 12:10 Bug Detection and Localization using Pretrained Code Language Models Viola Campos
12:10 - 12:30 Evaluating Contextualized Code Search in Practical User Studies Johannes Villmow
- lunch break -
Session "LLMs and Forecasting" Markus Schröder / Dirk Krechel (chairs)
14:00 - 14:20 Leveraging Distillation Techniques for Document Understanding: A Case Study with FLAN-T5 Marcel Lamott
14:20 - 14:40 Computer-Assisted Short Answer Grading Using Large Language Models and Rubrics Tim Metzler
14:40 - 15:00 Anomaly Detection in Supermarket Refrigeration Systems using Transformer Models: A Comparative Study Melina Meyer
15:00 - 15:20 Revealing the Effects of Data Heterogeneity in Federated Learning Regression Models for Short-Term Solar Power Forecasting Marc Haller
- closing and coffee -

Call for Papers

Recent technological successes of artificial intelligence have become the subject of many discussions. Publicly and often, AI methods are claimed to have a transformative impact on our working world: AI can support routine tasks, help us gather, share and reuse knowledge, analyze complex data patterns highly efficiently and at enormous scale, serve as a smart assistant, discover and prevent failure situations and fraud, etc. This way, AI technology can make an important contribution towards an efficient economy in times of shortage of qualified workforce. Due to the domain-agnostic nature of machine learning, AI affects various sectors, for example IT and engineering, the educational sector, logistics, and health care.

The AI@WORK workshop 2024 brings together AI researchers and practitioners interested in innovative AI applications for working environments across all sectors. This includes, but is not limited to, applications in industrial scenarios (such as smart manufacturing and maintenance) over health care (such as data-driven diagnosis and clinical care) to knowledge-intensive tasks (such as information retrieval and teaching).

We encourage research contributions related to transformative practical applications of AI in work places across all sectors. Particularly, we encourage submissions on (but not limited to) the following topics:

  • innovative AI models
    such as (virtual) knowledge graphs, vector databases, embeddings, or cross-domain learning
  • innovative AI applications
    such as LLM-based IR and QA, AI-supported educational scenarios (teaching, training, grading, feedback), or AI-supported workflows and decision processes
  • interactive AI systems
    such as hybrid teams of human and AI agents, context-sensitive, or neurosymbolic support of knowledge work
  • limitations of AI methods (e.g., defects in AI-generated content, resource efficiency and performance aspects) and trustworthiness (e.g., data leakage, privacy, safety, transparency).

Review Process

We kindly invite workshop contributions in a two-stage process:

  • First, abstracts (in PDF) of up to 3,000 characters must be submitted. These will be reviewed and selected by our program committee, and selected submissions will be presented in the workshop at INFORMATIK Festival 2024 in Wiesbaden.
  • Second, all presenters will also be invited to submit a final paper for the workshop proceedings of INFORMATIK 2024. Final papers will again be peer-reviewed by the program committee before they are accepted for publication. Accepted papers which will be published digitally in the GI’s public library (you can find more information here). Uploading your paper on arXiv is possible (see the granting of rights form).
The workshop itself will follow a highly interactive format, and is planned to include both oral presentations and a poster session. For oral presentations, ample time for discussion will be reserved.

Important Dates

  • Apr 10: Submission deadline (one-page abstract)
  • Apr 15: Notification of acceptance (one-page abstract)
  • May 20 May 27: Submission deadline (final paper)
  • Jun 05 Jun 11: Notification of acceptance (final paper)
  • Jun 27: Camera-ready submission
  • Sep 24: Workshop date
All deadlines Berlin time, 23:59.

Formatting & Page Limit

Your abstract should come in PDF format, be in English, not exceed 3,000 characters, and include the title, all authors, their affiliations, and contact information (there is no template).

Full papers should be limited to 8 pages content + max. 3 pages of references/appendix and follow the GI author guidelines. Do adhere to the LaTeX template (as zip-file or on overleaf). Again, do include all authors, affiliation, and contact information.

Submission

To upload your abstracts and papers, please use the submission system (Microsoft CMT).

Registration

Tickets can be purchased via the INFORMATIK Festival's registration page.

Please note that all workshop participants require a ticket.

Also, for each accepted paper to be published, at least one author must present a ticket. It is advisable to choose the person who will ultimately present the paper on site. The invoice number of the ticket must be included in the full paper submission.

Contact

You'll reach us best via e-mail ().

Also, feel free to contact Adrian Ulges.

Organization Committee


We thank the following reviewers for their support:

  • Mahta Bakhshizadeh, DFKI
  • Desiree Heim, DFKI
  • Michael Schulze, DFKI