Automated Machine Learning Lab

Current Research Topics

Metalearning and Automatic Model Recommendation Based on Knowledge-base
Publications List:
  1. Automatic Machine Learning Derived from Scholarly Big Data

  2. A Hybrid Approach for Automatic Model Recommendation

  3. AutoGRD: Model Recommendation Through Graphical Dataset Representation

Meta-Features and Auto Feature Selection\Generation
Publications List:
  1. ExploreKit: Automatic Feature Generation and Selection

Optimal Machine Learning Pipeline Search and Recommendation 
Publications List:
  1. RankML: a Meta Learning-Based Approach for Pre-Ranking Machine Learning Pipelines

  2. DeepLine: AutoML Tool for Pipelines Generation using Deep Reinforcement Learning and Hierarchical Actions Filtering

Explainability of the Models\Instances 
  1. Explaining anomalies detected by autoencoders using SHAP

Neural Network Architecture Search 
  1. Source Model Selection for Deep Learning in the Time Series Domain

  2. A meta-learning Approach for Image Classification Architecture Recommendation

Additional Research Topics
  1. Model Robustness (general and over time)

  2. Data Augmentation

  3. Learning on Budget

  4. Model Compression

Lab Research Team