Results of the Engineer 4 Science 2022 Competition - results of the competition for the Field of Information Technology

Opublikowano: 09/06/2022 1:25 pm

IEEE logo small

Winning entry (goes to the next stage – the national competition - Superstar 4 Science 2022)

  • Job Title: Image Captioning with Deep Learning
  • Authors: Wiktor Łazarski, Eng.
  • Supervisor: Rajmund Kożuszek, MSc Eng.

The work concerns the automatic description of images. It combines the understanding of images with natural language modeling, both syntactically and semantically. The author used the mechanism of attention to determine the relationships between the semantic features of the image and the words. The proposed configuration of cooperating neural networks and the procedure for teaching them allowed to obtain a solution with high efficiency. As part of his master's thesis, Eng. Wiktor Łazarski deals with the segmentation of eye components visible in infrared images. It will be part of a biometric system using iris recognition.
First Degree Honor

  •  Job Title:Extending a workflow engine for bioinformatics pipelines execution in a distributed environment
  • Authors: Dominika Maziec Eng.
  • Supervisor: Agnieszka Szmurło,

MSc Eng. Dominika Maziec worked in the area of application of Big Data technology in genomic data processing. In the awarded work, the author made a critical analysis of existing solutions used in medical centers, drawing attention to their insufficient efficiency and lack of automation. It proposed and implemented an automated, monitorable processing pipeline using a recognized workflow platform. By adapting the tools for use in a distributed environment, the author ensured the scalability of the solution, which was confirmed by performance tests. Dominika Maziec successfully continues her development in the field of bioinformatics. Even before defending her engineering thesis, she collaborated on a project for the Children's Health Center and currently works as a bioinformatics engineer at Harvard Medical School.
Second degree distinction

  • Job Title: Efficient implementation of a graph neural network
  • Authors: Jakub Grzechociński, Eng.
  • Supervisor: Paweł Wawrzyński, PhD, Eng.

Graph neural networks are networks that operate on data in the form of graphs. Networks of this type are used in many fields and solve various problems, such as classification, edge replenishment or graph generation. The author has implemented a graph neural network with an autocoder architecture. The aim of the implementation was to create a competitive solution that is able to process graph data of various sizes, allows the architecture to be adapted to many tasks and data sets, and is computationally efficient due to the capabilities of GPUs. The architectures and technologies used in this implementation and in the solutions of other authors were presented. The concept of network operation, methods of effective presentation of it in the code and a summary of the results were described.All winners will be honored with the prestigious IEEE Diploma. The winning works and first-degree distinctions additionally receive material prizes. Congratulations to the authors of the works and their Guardians!Link to the announcement of the Competition Organizer: https://results.e4s2022.4scienceinstitute.org/  
The participant of the "Engineer 4 Science 2022" competition may be a Graduate who has obtained the title of Engineer, and his diploma thesis has been evaluated by the Promoter and reviewer for two very good grades. In the process of evaluating works, the Jurors are guided by the following criteria:

  • timeliness of the problem,
  • innovativeness
  • idea for the implementation of the topic,
  • substantive value,
  • the level of performance of the practical part.

The works of the Laureates of local competitions "Engineer 4 Science 2022" are automatically qualified for the Super Star 4 Science 2022 Competition and for thematic Competitions held at the national level.
The "Engineer 4 Science" competition is a cyclical event organized by 4 Science Institute Sp. z o. o. The patron of "Engineer 4 Science 2022" is PS IEEE (Polish Section of the Institute of Electrical and Electronics Engineers), GovTech Polska and the Association of Electronics Graduates at the Warsaw University of Technology @ELKApw. The Business Partner and the founder of the Competition awards is the company: AssisTech, Roche Global IT Solution Centre, Philips. The Technological Partner of the Competition, which provides Digital Diplomas for Laureates based on Blockchain technology, is Billon Sp. z o.o.