Schedule

 See accepted papers for all details on the invited talks and papers.



 08:30 - 08:40  Welcome
 08:40 - 09:20  Invited talk 1: Luc de Raedt
 09:20 - 09:40 Contributed talk 1: AlphaD3M: Machine Learning Pipeline Synthesis
 09:40 - 10:40 Contributed talk 2: Towards AutoML with Concept Drift: First Results
 10:00 - 10:10 Poster spotlights 1
  • Annotative Experts for Hyperparameter Selection
  • Adaptive Quadrature for Fast Sequential Hyperparameter Marginalization
  • Clustered gradient-based meta-learning for heterogeneous and evolving tasks
  • ML-Plan for Unlimited-Length Machine Learning Pipelines
  • MOSAIC: AutoML with Monte Carlo Tree Search
  • Workflow Recommendation for Text Classification with Active Testing
  • Ensembles with Automated Feature Engineering
  • CapsAndRuns: An Improved Method for Approximately Optimal Algorithm Configuration
  • Demo paper: Far-HO: A Bilevel Programming Package for Hyperparameter Optimization and Meta-Learning
  • Demo paper: Tune: A Research Platform for Distributed Model Selection and Training
 10:10 - 10:30 Poster session 1: morning posters (papers shown in poster spotlights 1 & 2)
 10:30 - 11:00
 Coffee Break
 11:00 - 11:40
 Invited Talk 2: Zoubin Ghahramani
Automating machine learning
 11:40 - 11:50 Poster spotlights 2
  • Bayesian Optimization Exploiting Monotonicity and Its Application in Machine Learning Hyperparameter Tuning
  • Dealing with Categorical and Integer-valued Variables in Bayesian Optimization with Gaussian Processes
  • P4ML: A Phased Performance-Based Pipeline Planner for Automated Machine Learning
  • Scalable Meta-Learning for Bayesian Optimization using Ranking-Weighted Gaussian Process Ensembles
  • Dynamic Network Architectures
  • Automatic Feature Selection in Learning Using Privileged Information
  • Neural Network Conversion of Machine Learning Pipelines
  • Backprop Evolution
  • Demo paper: Orion: Experiment Version Control for Efficient Hyperparameter Optimization
  • Demo paper: AMLA: an AutoML frAmework for Neural Network Design
 11:50 - 12:30  Poster session 2: morning posters (papers shown in poster spotlights 1 & 2)
 12:30 - 14:00  Lunch Break
 14:00 - 14:20  Contributed Talk 3: Best Paper Award Meta-Learning Transferable Active Learning Policies by Deep Reinforcement Learning
 14:20 - 14:35  Poster spotlights 3:
  • Automated Machine Learning for Soft Voting in an Ensemble of Tree-based Classifiers
  • Practical Automated Machine Learning for the AutoML Challenge 2018
  • Towards Further Automation in AutoML
  • Predicting hyperparameters from meta-features in binary classification problems
  • Decision-Theoretic Meta-Learning: Versatile and Efficient Amortization of Few-Shot Learning
  • Abstractive Tabular Dataset Summarization via Knowledge Base Semantic Embeddings
  • Preprocessor Selection for Machine Learning Pipelines
  • Towards Automated Deep Learning: Efficient Joint Neural Architecture and Hyperparameter Search
  • Automatic Type Inference with a Nested Latent Variable Model
  • A Comparison of Several Models on Hardware Accelerator Data for Deep Neural Networks Design using Bayesian Optimization
  • Similarity encoding for learning with dirty categorical data
  • Demo paper: Distil: A Mixed-Initiative Model Discovery System for Subject Matter Experts
  • Demo paper: Automatic Gradient Boosting
  • Demo paper: Efficient Neural Architecture Search with Network Morphism
 14:35 - 15:30 Poster session 3: afternoon posters (papers shown in poster spotlight session 3, and contributed talks)
 15:30 - 16:00  Coffee Break
 16:00 - 16:40 Invited Talk 3: Chelsea Finn Properties of a Good Meta-Learning Algorithm (and How to Achieve Them)
 16:40 - 17:00 Contributed talk 4: Evolutionary Algorithms and Reinforcement Learning: A Comparative Case Study for Architecture Search
17:00 - 18:00  Community/Panel Discussion


Comments