Author Archives: Mikko J Raatikainen

Two awards for ESE researchers

|

 Juhani Kivimäki and Mikko Raatikainen were recently awarded for their research. 

Juhani Kivimäki received an award for the best master’s thesis of the year 2023 in Computer Science in Finland from the Finnish Society for Computer Science with the title “Uncertainty Estimation with Calibrated Confidence Scores”. The thesis was carried out in the IML4E project (iml4e.org) funded by Business Finland in collaboration with Finnish company Basware and supervised by Professor Jukka K. Nurminen and D.Sc. Mikko Raatikainen.  

Mikko Raatikainen received in IEEE International Conference on Software Architecture (ICSA’24) the Most Influential Paper award, papers from CBSE 2014, WICSA 2014, and QOSA 2014. ICSA is a top-level conference (Jufo 2 in the Finnish classification system) that merged these other three conferences. The paper is a joint work by Vittorio Dal Bianco, Varvana Myllärniemi, Marko Komssi, and Mikko Raatikainen titled as “The role of platform boundary resources in software ecosystems: a case study.”  

Further information: 

Publication in ICSA conference

|

We presented last week a paper about the industrial experience of micro frontends at the international conference on software architecture (ICSA). The conference is one of the most high-quality conferences in software engineering and rated as Jufo2 in the Finnish ranking system, i.e., the highest level for conferences.

See the linkedin post (Finnish) and the research portal for the paper and presentation.

 

 

 

 

 

 

New demo video: Model card validation

|

We recently published a new demo video demonstrating the key features of the model card validation action developed at the University of Helsinki.

The ultimate goal is to ensure that every model promoted from the experimentation phase to production is accompanied by an up-to-date model card. This tool helps by adding an automated testing step (the validation action) to an MLOps pipeline, where the existence, structural integrity, and syntactical validity of the model card are confirmed.

Key features demonstrated:

  • Validation of the model card
  • Creating custom rules
  • Creating custom functions for the rules
  • Using the MLflow integration

Source code: https://github.com/CompliancePal/mode…

More information on usage: https://modelcards.compliancepal.eu/

Check out our latest AI/ML-systems and quantum computing results for FCAI AI day.

|

ESE group will present the recent research results in FCAI day on November 16, 2022. See the following contributions that will be there:

Talks:

  • Lalli Myllyaho, Jukka K. Nurminen and Tommi Mikkonen: Node co-activations as a means of error detection—Towards fault-tolerant neural networks
  • Kai-Kristian Kemell and Ville Vakkuri: ECCOLA – A method for implementing ethically aligned AI systems

Posters:

  • Ilmo Salmenperä and Jukka K. Nurminen: Hyperparameter Optimization of Quantum Annealing Boosted Restricted Boltzmann Machines
  • Juha Mylläri: Augmenting the Student-Teacher Feature Pyramid Matching Method for Better Unsupervised Anomaly Localization
  • Juhani Kivimäki: Uncertainty Estimation with Calibrated Confidence Scores
  • Jussi Steenari, Lucy Ellen Lwakatare, Jukka K. Nurminen, Jaakko Talonen and Teemu Manderbacka: Mining port operation information from AIS data
  • Yumo Luo, Mikko Raatikainen, Lalli Myllyaho and Jukka K. Nurminen: From Misbehaviour and Fault Tolerance in ML system towards dependable and self-improving MLOps (edited) 

 

The final seminar of the AIGA (AI governance) project 13.10.2022 Online&Turku

|

Artificial Intelligence Governance and Auditing (AIGA) project, which ESE is a participant, will have its Final Seminar. We will present the key results of the project. We will also host more general discussions about the latest developments around responsible AI. Both researchers and industry representatives will share their views in the open seminar, and the program ends with an informal networking event.

The final seminar will take place on 13.10.2022 at the Turku School of Economics. A live stream is provided for remote participation. Please make sure to register before 5.10. (16:00).

For further information, please visit our event website: https://ai-governance.eu/final-seminar/

Register as a participant: https://konsta.utu.fi/Default.aspx?tabid=88&tap=13197

MLOps research and thesis positions

|

Machine learning (ML) is becoming an increasingly common technology that is used in software-based systems. While software and software engineering practices are still a quintessential part of the design of such systems, the design space has data and ML and respective data scientist stakeholders as new concerns. Hence, MLOps — as a derivative of DevOps — is currently emerging to better integrate data and ML into the software engineering way of working, as well as integrate ML better in the software-based systems. MLOps is about both practices and tools for ML-based systems that technically enable iterative software engineering practice.

The ESE research group is strengthening its research focus in the area of MLOps. In particular, we are participating in projects that study MLOps, such as AIGA https://ai-governance.eu/ and recently started IMLE4 https://itea4.org/project/iml4e.html.

We have funded thesis positions in the area of MLOps in these research projects that can be tailored to the interest of the applicant. For details, contact Mikko Raatikainen (first.last@helsinki.fi).