Research

The research interests of the mlearn group cover a variety of subfields where machine learning meets data science, intelligent systems, knowledge representation and reasoning, multimedia and multimodal content analysis, indexing and retrieval, semantic metadata interoperability and search, behaviour recognition and affective interactions, sentiment analysis, as well as applications in medical data analysis and health care monitoring, cultural content search and digital libraries, computational finance, computational advertising, smart homes, smart cities and Internet of Things.

Some highlights of these research interests are listed below:

example_label2                                          Optical character Verification in food labels

radial_distribution

Mlearn1

 

 

Machine learning in signal processing domain for analysing nuclear data

 

 

Machine learning is one of the basic technologies in data science and big data analysis

 

 

Machine learning coupled with reasoning technologies is the core of  systems that:  handle data from various sources; take into account domain knowledge; related data from external links; interact with the users.

 

Machine learning coupled with reasoning technologies is the core of systems that handle data from various sources; take into account domain knowledge; related data from external links; and interact with users.

 

 

mlearn4

 

Interweaving intelligent technologies and semantic web principles can provide systems that analyse multimedia content with knowledge re-use capability and adaptability

 

 

 

 

 

mlearn5

Visual image retrieval & visualisation; viral.image.ntua.gr

 

 

 

 

 

 

 

 

mlearn6

Computational advertising has emerged as a new sub-discipline in computer science, bridging the gap among the areas such as information retrieval, data mining, machine learning, economics, and game theory.

Research in computational advertising includes: mechanism design for online auctions and guaranteed system; learning algorithms for optimal bidding strategies; mining user effects and engagement.

 

 

mlearn7Applications of machine learning in finance can bring automation and solve the limitations of the existing financial models.

Research directions include: financial time series prediction and pattern recognition; optimal trading strategies for statistical arbitrage; intelligent fraud detection for risk management; and customers’ behaviour analysis in consumer finance.

 

 

Machine learning is a basic technology used in affective computing; particularly in behavioural analysis in human machine interaction

Machine learning is a basic technology used in affective computing; particularly in behavioural analysis in human machine interaction

 

 

 

 

 

 

mlearn9

 

Generating robots that learn to recognise users’ actions towards them and react coherently.

 

 

 

mlearn91

 

EUROPEANA: The European Digital Library; a single point of entry to over 50 millions of European cultural objects

 

Visual search of European cultural  heritage (view.image.ntua.gr)

Visual search of European cultural
heritage (view.image.ntua.gr)

 

 

 

 

 

 

 

 

Presentation at USA National Archives of the awarded MINT system (Washington DC, 2011)

Presentation at USA National Archives of the awarded MINT system (Washington DC, 2011.)

 

 

 

 

Leave a Reply

Your email address will not be published. Required fields are marked *