Computers Don’t Byte: XAI
podcast, LIACS podcast, Online
🚢⚙️ Exciting News: “Computers Don’t Byte” Podcast by LIACS! 🎙️
Talks and presentations
DoE2Vec: Deep-learning Based Features for Exploratory Landscape Analysis
poster, GECCO 2023, Lisbon, Portugal
Deep-Bias: Detecting Structural Bias using Explainable AI
poster, GECCO 2023, Lisbon, Portugal
BIAS: A Toolbox for Benchmarking Structural Bias in the Continuous Domain
talk, GECCO 2023, Lisbon, Portugal
Hot off the Press presentation @Gecco.
AI for Expensive Optimization Problems in Industry
poster, CAI 2023, Santa Clara, California, United States
The optimization of real-world engineering problems can be a challenging task, due to the limited understanding of problem characteristics and the high computational cost of objectives and constraints. This study proposes an AI-assisted optimization pipeline that addresses these challenges by using proxy functions in order to select and optimize an optimization algorithm and its hyper-parameters. It thereby significantly accelerates the optimization process on the real (expensive) problem. To obtain such proxy functions Exploratory Landscape Analysis (ELA) features are used to characterize the problem’s landscape. The ELA features are then used to identify an artificial function that replicates the original problem’s properties.
Functional Kernel Hypothesis Testing for Channel Selection in Time Series Classification
poster, CAI 2023, Santa Clara, California, United States
Multi-channel time-series classification is a challenging task. With sometimes thousands of sensors available for real-wolrd applications, it is a daunting and difficult task to select which channels to include and which not.
Neural Architecture Search VS Black Box Optimization
Invited talk, BMW invited presentation, Munich, Germany
Invited presentation at BMW headquarters
Neural Network Design: Learning from Neural Architecture Search
Talk, SSCI 2020, Canberra, Australia
Automated Deep Learning
Invited talk, SAILS 2020,
Invited presentation at SAILS (online)
A New Approach Towards the Combined Algorithm Selection and Hyper-parameter Optimization Problem
Talk, SSCI 2019, Xiamen, China
Deep Learning, A broad introduction
Invited talk, Ecole Summer school, Leiden, The Netherlands
Invited talk for the Summer school for Early Stage Researchers. This summer school takes place within the EU project Experience-based Computation: Learning to Optimise (ECOLE), which investigates novel synergies between nature-inspired optimisation and machine learning to address key challenges faced by the European industry.
Automatic Configuration of Deep Neural Networks with Parallel Efficient Global Optimization
Talk, IJCNN 2019, Budapest, Hungary
Deep Steel: Optimizing networks to discover defects on steel surface
Invited talk, Tata Steel, IJmuiden, The Netherlands
Invited talk at Tata Steel, about using deep learning to discover steel surface defects.
Deep Steel: Optimizing networks to discover defects on steel surface
Invited talk, ECOLE workshop, Leiden, The Netherlands
Invited talk for the ECOLE program, learning to optimize.
A Novel Uncertainty Quantification Method for Efficient Global Optimization
Talk, IPMU 2018, Cadiz, Spain
From Data to Knowledge to Optimization
Invited talk, PhD Colloquium, Leiden, The Nethehrlands
Invited talk for thhe PhH Colloquium with Math and Informatics
Local Subspace-Based Outlier Detection using Global Neighbourhoods
Talk, IEEE Big Data, Washington DC, USA
Analysis and Visualization of Missing Value Patterns
Talk, IPMU 2016, Eindhoven, The Netherlands
An Incremental Algorithm for Repairing Training Sets with Missing Values
Talk, IPMU 2016, Eindhoven, The Netherlands
Datamining & Apps
Invited guest lecture, Data Science course, Liacs, Leiden, The Netherlands
Invited lecture at the Data Science course, LIACS, Leiden University
Fuzzy Clustering for Optimally Weighted Cluster Kriging
Talk, WCCI 2016, Vancouver, Canada
Fuzzy Clustering for Optimally Weighted Cluster Kriging
Invited talk, PhD Seminar, Liacs, Leiden, The Netherlands
Invited talk for the PhD Seminar at LIACS, Leiden University
Optimally Weighted Cluster Kriging
talk, IDA 2015, Saint-Etienne, France
In business and academia we are continuously trying to model and analyze complex processes in order to gain insight and optimize. One of the most popular modeling algorithms is Kriging, or Gaussian Processes. A major bottleneck with Kriging is the amount of processing time of at least O(n3) and memory required O(n2) when applying this algorithm on medium to big data sets. With big data sets, that are more and more available these days, Kriging is not computationally feasible. As a solution to this problem we introduce a hybrid approach in which a number of Kriging models built on disjoint subsets of the data are properly weighted for the predictions. The proposed model is both in processing time and memory much more efficient than standard Global Kriging and performs equally well in terms of accuracy. The proposed algorithm is better scalable, and well suited for parallelization.
Datamining & Apps
Invited talk at the CWI, Amsterdam