Deep Learning: a tech break through

Activity: Talk or presentationInvited talkunknown

Description

Deep learning has emerged as a highly successful machine learning approach and has already been impacting a wide range of fields. MIT Technology Review selected it as one of the 10 tech break throughs in 2013. Deep learning is founded on novel algorithms and neural network architectures in machine learning together with recent availability of GPU and massive data. In its core, deep learning discovers multiple levels of distributed representations of the input, with higher levels representing more abstract concepts. These representations led to impressive successes in computer vision, speech recognition, and Internet advertising. IT giants like Google, facebook, Microsoft, Baidu consider Deep Learning as the key technology of the future. I will show some of our Deep Learning applications. We used Deep Learning for toxicity prediction, for biclustering, and for vision tasks. Biclustering identified customer groups together with product groups, that is, particular customers that select a specific set of products. Using unsupervised rooting we achieved state-of-the-art performance on several vision benchmark data sets. Finally, I will show how LSTM networks that were developed by us are currently used for language and text processing at Google, Microsoft, and Facebook.
Period21 May 2015
Event titleunbekannt/unknown
Event typeOther
LocationGermanyShow on map

Fields of science

  • 305 Other Human Medicine, Health Sciences
  • 102019 Machine learning
  • 304 Medical Biotechnology
  • 303 Health Sciences
  • 302 Clinical Medicine
  • 301 Medical-Theoretical Sciences, Pharmacy
  • 102 Computer Sciences
  • 106005 Bioinformatics
  • 106007 Biostatistics
  • 304003 Genetic engineering
  • 106041 Structural biology
  • 102010 Database systems
  • 101018 Statistics
  • 106023 Molecular biology
  • 106002 Biochemistry
  • 102001 Artificial intelligence
  • 102015 Information systems
  • 101004 Biomathematics
  • 102004 Bioinformatics

JKU Focus areas

  • Health System Research
  • Computation in Informatics and Mathematics
  • Clinical Research on Aging
  • Nano-, Bio- and Polymer-Systems: From Structure to Function
  • Medical Sciences (in general)