Alfavita Newsroom Predictive analytics is increasingly used in industry, business, finance, government, healthcare, and education, using methods from machine learning, data mining and statistics. The use of predictive analytics can lead to improved and efficient decision making but the models that are generated are often highly complex. This makes it hard to understand, trust, and judge the results of such models for humans.
The project 'Transparent Explanations for Predictive Analytics using Interactive Visualization', funded by an NWO-TOP grant, attacks this challenge. We put the human central, and strongly believe that close collaboration between experts on visual analytics, cognitive psychology, machine learning, and human machine interaction is needed to make major steps forward. We aim at building a team of three PhD students with complementary expertise and interests, supervised by staff members from our data mining, visualization, and human-technology interaction groups. Specifically, we have the following positions available:
(A) PhD student Predictive Analytics A complex model cannot be understood by just looking at all structures and parameters involved. Typical questions are what features were used and how important these were, in general as well as in specific cases. Complex models can be transformed into quantitatively or qualitatively simpler models. A key challenge is to do this effectively and efficiently, such that the simplified model is still relevant. Quantification of the scope and limitations are important to guide the human viewer. This PhD student will be supervised by prof.dr. Mykola Pechenizkiy and prof.dr. Maurits Kaptein.
(B) PhD Interactive Visualization Visualization will be used to present explanations: the reasoning used, uncertainty involved, relevance of variables, similar cases, etc. One challenge is that the audience of model users is varied, ranging from data experts via domain experts to lay users, each with different requirements and interests. Ultimately, we aim to enable them to navigate smoothly through complex models, guided by their interests, showing the big picture, as well as detail when needed. This PhD student will be supervised by prof.dr. Jack van Wijk and dr. Michel Westenberg.
(C) PhD Human Model Interaction All novel solutions have to be carefully evaluated, for effectiveness, efficiency, and usability. This is notoriously hard in visual analytics, and requires a deeper understanding of fuzzy concepts such as understandability, trust, and acceptance. We aim at the development of models and benchmarks for this, as well as thorough evaluation in real-world cases. This PhD student will be supervised by prof.dr. Wijnand IJsselsteijn, prof.dr. Chris Snijders, and dr. Martijn Willemsen.
The PhD students will be appointed at the Department of Mathematics and Computer Science (PhD A and B) and the Department of Industrial Engineering & Innovation Sciences (PhD C) of TU/e. The project is the result of cooperation within the Jheronimus Academy of Data Science (JADS): a unique collaboration of two renowned universities, TU/e and Tilburg University, bringing together their top academics to further the emerging and multidisciplinary academic field of data science. The project team will work together at our new JADS Mariënburg campus in 's-Hertogenbosch, which is fully dedicated to data science.
Για να δείτε περισσότερες πληροφορίες πατήστε ΕΔΩ
Όλες οι σημαντικές και έκτακτες ειδήσεις σήμερα
ΕΛΜΕΠΑ: Το κορυφαίο πρόγραμμα Ειδικής Αγωγής στην Ελλάδα για διπλή μοριοδότηση
Το 1ο στην Ελλάδα Πρόγραμμα επιμόρφωσης Τεχνητής Νοημοσύνης για εκπαιδευτικούς με Πιστοποιητικό
ΑΣΕΠ: Η πιο Εύκολη Πιστοποίηση Αγγλικών για μόρια σε 2 ημέρες (δίνεις από το σπίτι σου με 95 ευρώ)
Παν.Πατρών: Μοριοδοτούμενο σεμινάριο ΕΙΔΙΚΗ ΑΓΩΓΗΣ με 65Є εγγραφή - έως 25/11
ΕΥΚΟΛΕΣ πιστοποιήσεις ΙΣΠΑΝΙΚΩΝ - ΙΤΑΛΙΚΩΝ - ΓΑΛΛΙΚΩΝ - ΓΕΡΜΑΝΙΚΩΝ για ΑΣΕΠ - Πάρτε τις ΑΜΕΣΑ
2ος Πανελλήνιος Γραπτός Διαγωνισμός ΑΣΕΠ: Τα 2 μαθήματα εξέτασης και η ύλη
