Topics of Research

  • machine learning and data mining,
  • modelling of numerical, symbolic, and image data,
  • image analysis,
  • medical diagnostics and prognostics,
  • estimation of data quality and data importance,
  • statistical data analysis,
  • deep learning,
  • interactions of various parameters,
  • text summarization,
  • web users profiling,
  • network analysis,
  • text mining,
  • graph mining.

Machine learning and data mining search for regularities in moderate to large databases by learning models of data.  Generated models can be used for explanation of data, simulations, process control, prediction, and solving new related problems. An example is medical diagnostics, where from previously diagnosed patients a model for diagnosing new, previously unseen patients can be derived. Similar approaches can also be used in insurance or banking businesses, where predictive models can be used for detecting unusual or interesting patterns in a day-to-day business process.

Infrastructure

Besides desktop computers and notebooks, the laboratory is equipped with the following powerful computers:

1) General purpose server - Dell PowerEdge M630 (Contact in LKM: Marko Robnik Šikonja)

  • CPU: 2 x Intel Xeon E5-2630 V3 / 8 core, 2.40-3.20 GHz, 20 MB
  • Total number of CPU: 16 (32 - Hyperthreading)
  • RAM: 384 GB DDR4 2400 MHz
  • HDD: 22 TB

2) GPU computing machine (Contact in LKM: Tadej Škvorc)

  • CPU: Intel i7-7820X / 8 core, 3.60-4.30 GHz, 11 MB
  • Total number of CPU: 8 (16 - Hyperthreading)
  • GPU: 2 x NVIDIA GeForce GTX 1080 Ti / 11 GB GDDR5X
  • Total GPU memory: 22 GB
  • RAM: 32 GB DDR4 2400 MHz
  • HDD: 4 TB + 500 GB SSD

3) Machine for Bayesian computing (Contact in LKM: Erik Štrumbelj)

  • CPU: AMD Ryzen Threadripper 1950X / 16 core, 3.40-4.00 GHz, 32 MB
  • Total number of CPU: 16 (32 - Hyperthreading)
  • RAM: 64 GB DDR4 3200 MHz
  • HDD: 512 GB SSD


Survey of activities

Here you can download our yearly surveys: