research areas explored by the machine learning and inference laboratory

MLI conducts basic and applied research in the areas of machine learning and intelligent systems. Currently MLI is involved in several research projects which focus on handling heterogeneous/complex data sources, background knowledge and semantics as well as transparency and understandability of results. Our efforts are on theoretical, algorithmic, and application developments. Among the important current projects are (list is not inclusive):

Machine Learning and Inference Reporting Criteria for Transparency and Reproducibility

COVID-19 Social Distancing and Symptom Modeling

WiFi Contact Prediction

– Ontology-guided Machine Learning

– Intelligent Patient Data Generator

– Learning from aggregated data and published results

– Individualized comparative effectiveness of medical treatments

– Machine learning-guided evolutionary optimization for engineering design

– Learning from patient’ medical records

– Predicting Patients’ Functional Status

GPS-based Analysis of Wandering Patterns of Individuals with Alzheimer’s

Over the past 40 years, the laboratory originated several research areas, and participated in many others. Selected important areas of research, along with main researchers that worked on these topics are listed below.

Theories of Learning, Inference, and Discovery
Machine Learning Systems
Data Mining and Knowledge Discovery
Healthcare Applications of Machine Learning
User Modeling and Intrusion Detection
Non-Darwinian Evolutionary Computation: Learnable Evolution Model
Machine Vision through Learning
Computational Intelligence and Machine Learning Virtual Community

Theories of Learning, Inference, and Discovery

Inferential Theory of Learning (Michalski, Sklar, Bloedorn, Kaufman)
Multistrategy Task-Adaptive Learning: MTL (Michalski, Wnek, Kaufman, Zhang)
Knowledge Representation Using Dynamically Interlaced Hierarchies  (Michalski, Alkharouf)
Cognitive Models of Plausible Reasoning (Michalski, Sklar)
Learning Goals in Multistrategy Learning (Michalski, Utz)

Machine Learning Systems

Environments for Natural Induction: STAR-AQ19/AQ21(Michalski, Kaufman, Wojtusiak, Bloedorn, Fischthal)
Learning Multihead Attributional Rules (Michalski, Pietrzykowski, Glowinski)
Data-driven Constructive Induction: AQ17-DCI (Michalski, Wojtusiak, Bloedorn)
Advice-driven Constructive Induction (Michalski, Wojtusiak)
Research Areas (Michalski, Wnek)
Multistrategy Constructive Induction: AQ17-MCI (Michalski, Bloedorn, Wnek)
Constructive Induction in Engineering Design  (Arciszewski, Michalski, Wnek, Bloedorn)
Constructive Induction Approach to Growing Neural Networks  (Wnek)

Knowledge Mining, Data Mining and Knowledge Discovery

Inductive Databases and Knowledge Scouts (Michalski, Kaufman, Pietrzykowski, Sniezynski, Wojtusiak, Seeman, Fischthal, Alkharouf, White, Draminski, Glowinski)
A Diagrammatic Visualization of Data Mining and Machine Learning Processes (KV) (KV) (Michalski, Szymacha, Sniezynski, Vang, Zhang, Wnek)
Knowledge Discovery in Databases: INLEN (Michalski, Kaufman, Ternstedt, Bloedorn, Kerschberg, Wnek)
Learning Problem-Oriented Decision Structures from Decision Rules(Michalski, Imam)
Expert Systems with Learning Capabilities (Michalski, Kaufman, Imam, Ribeiro)
Significance Vector Approach to Analysis of Ultra-Large Databases (Michalski, Goshorn)
Symbolic Meta-analysis of Aggregated Data (Wojtusiak, Baranova, Michalski, Simanivanh)
Learning Capabilities in Autonomous Transportation and Logistics (Wojtusiak, Michalski, Gehrke, Herzog)

Healthcare Applications of Machine Learning

Learning from Published Medical Results (Wojtusiak, Baranova, Michalski, Simanivanh)
Learning Natural Language Descriptions (Michalski, Wojtusiak)
Comparative Effectiveness Research Using Machine Learning (Wojtusiak)
Automated Advising in Machine Learning for Healthcare Professionals (Wojtusiak)

User Modeling and Intrusion Detection

Learning User Signatures through Symbolic Learning (LUS) (Michalski, Kaufman, Pietrzykowski, Wojtusiak)

Non-Darwinian Evolutionary Computation: Learnable Evolution Model

Multistrategy Evolution using the Learnable Evolution Model (Michalski, Wojtusiak, Kaufman)
Learnable Evolution Model in Engineering Design (Michalski, Kaufman, Wojtusiak)

Machine Vision through Learning

Multi-Level Image Sampling and Interpretation: MIST Methodology  (Michalski, Duric, Zhang, Maloof)
Machine Vision and Learning  (Michalski, Duric, Maloof, Zhang, Wnek, Bloedorn) (with Computer Vision Laboratory of the University of Maryland at College Park, Rosenfeld, Aloimonos, Davis)
Multistrategy Learning Vision Tasks by Integrating Symbolic and Neural Net Learning for Vision Tasks (Michalski, Zhang)
Learning to Recognize Shapes  (Michalski, Duric, Maloof)
Dynamic Recognition (Michalski, Bloedorn)

Computational Intelligence and Machine Learning (CIML) Virtual Community

The Machine Learning and Inference Laboratory collaborates with the University of Louisville on the National Science Foundation supported project whose goal is to build a CIML virtual community.
As a part of the project we created an initial database of CIML resources on the web. The database is being extended with new entries, and methods for guiding users through the resources are being developed.


Integrated Learning Systems for Education and Research: iAQ (Pietrzykowski, Michalski, Kaufman)
Integrated Learning Systems for Education and Research: Emerald JAVA Version (Michalski, Galkowski, Gorczyca, Jakubiec, Jura, Pawlowski, Rajda)
Integrated Learning Systems for Education and Research: Emerald Sun Version (Michalski, Kaufman, Bloedorn)