You are here: Home › Publications

Publications

2023

  • Arlind Kadra, Maciej Janowski, Martin Wistuba, Josif Grabocka
    Power Laws for Hyperparameter Optimization
    Thirty-Seventh Annual Conference on Neural Information Processing Systems (NeurIPS 2023)
  • Sebastian Pineda Arango, Josif Grabocka
    Deep Pipeline Embeddings for AutoML
    ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2023)
  • Abdus Khazi, Sebastian Pineda Arango, Josif Grabocka 
    Deep Ranking Ensembles for Hyperparameter Optimization
    International Conference on Learning Representations (ICLR 2023)
  • Gresa Shala, Thomas Elsken, Hadi Jomaa, Frank Hutter, Josif Grabocka 
    Transfer NAS with Meta-Learned Bayesian Surrogates
    International Conference on Learning Representations (ICLR 2023)
  • Gresa Shala, Andre Biedenkapp, Frank Hutter, Josif Grabocka 
    Gray-Box Gaussian Processes for Automated Reinforcement Learning
    International Conference on Learning Representations (ICLR 2023)

2022

  • Martin Wistuba, Arlind Kadra, Josif Grabocka (link)
    Supervising the Multi-Fidelity Race of Hyperparameter Configurations
    Thirty-Sixth Annual Conference on Neural Information Processing Systems (NeurIPS 2022)

  • Ekrem Öztürk, Fabio Ferreira, Hadi Jomaa, Josif Grabocka and Frank Hutter (link)
    Zero-shot AutoML with Pretrained Models
    International Conference on Machine Learning (ICML 2022)

  • Samuel Müller, Noah Hollmann, Sebastian Pineda Arango, Josif Grabocka, Frank Hutter (arXiv)
    Transformers Can Do Bayesian Inference
    International Conference on Learning Representations (ICLR 2022)

2021

  • Sebastian Pineda Arango, Hadi Samer Jomaa, Martin Wistuba, Josif Grabocka (arXiv)
    HPO-B: A Large-Scale Reproducible Benchmark for Black-Box HPO based on OpenML
    Proceedings of the Thirty-fifth Conference on Neural Information Processing Systems, Datasets and Benchmarks Track (NeurIPS 2021)

  • Arlind Kadra, Marius Lindauer, Frank Hutter, Josif Grabocka (arXiv)
    Well-tuned Simple Nets Excel on Tabular Datasets
    Proceedings of the Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS 2021)

  • Michael Ruchte, Josif Grabocka. (arXiv)
    Scalable Pareto Front Approximation for Deep Multi-Objective Learning.
    Proceedings of the IEEE International Conference on Data Mining (ICDM 2021).

  • Shayan Jawed, Hadi Jomaa, Lars Schmidt-Thieme and Josif Grabocka.
    Multi-task Learning Curve Forecasting Across Hyperparameter Configurations and Datasets.
    Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML 2021).

  • Ahmed Rashed, Lars Schmidt-Thieme and Josif Grabocka.
    A Guided Learning Approach for Item Recommendation via Surrogate Loss Learning.
    ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2021).

  • Wistuba, M., and Grabocka J. (2021), February. (pdf)
    Few-Shot Bayesian Optimization with Deep Kernel Surrogates.
    In the Ninth International Conference on Learning Representations (ICLR 2021).

2020

  • Jomaa, H. S., Schmidt-Thieme, L., & Grabocka, J. (2020). (arXiv)
    Dataset2vec: Learning dataset meta-features. 
    In Journal of Data Mining and Knowledge Discovery.
  • Jawed, S., Grabocka, J. and Schmidt-Thieme, L., 2020, May. (pdf)
    Self-supervised Learning for Semi-supervised Time Series Classification. 
    In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 499-511). Springer, Cham.
  • Drumond, R.R., Brinkmeyer, L., Grabocka, J. and Schmidt-Thieme, L., 2020. (pdf)
    HIDRA: Head Initialization across Dynamic targets for Robust Architectures.
    In Proceedings of the 2020 SIAM International Conference on Data Mining (pp. 397-405). Society for Industrial and Applied Mathematics.

2019

  • Rashed, A., Grabocka, J. and Schmidt-Thieme, L., 2019, September. (pdf)
    Attribute-aware non-linear co-embeddings of graph features.
    In Proceedings of the 13th ACM Conference on Recommender Systems (pp. 314-321).
  • Yalavarthi, V.K., Grabocka, J., Mandalapu, H. and Schmidt-Thieme, L., 2019, September. (pdf)
    Gait Verification using Deep Learning with a Pairwise Loss.
    In 2019 International Conference of the Biometrics Special Interest Group (BIOSIG) (pp. 1-7). IEEE.
  • Jameel, M., Grabocka, J. and Schmidt-Thieme, L., 2019, September. (pdf)
    Ring-Star: A Sparse Topology for Faster Model Averaging in Decentralized Parallel SGD.
    In Joint European Conference on Machine Learning and Knowledge Discovery in Databases (pp. 333-341). Springer, Cham.
  • Rashed, A., Jawed, S., Rehberg, J., Grabocka, J., Schmidt-Thieme, L. and Hintsches, A., 2019, September. (pdf)
    A Deep Multi-task Approach for Residual Value Forecasting.
    In Joint European Conference on Machine Learning and Knowledge Discovery in Databases (pp. 467-482). Springer, Cham.
  • Rashed, A., Grabocka, J. and Schmidt-Thieme, L., 2019, July. (pdf)
    Multi-Relational Classification via Bayesian Ranked Non-Linear Embeddings.
    In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 1132-1140).
  • Jomaa, H.S., Grabocka, J., Schmidt-Thieme, L. and Borek, A., 2019, July. (pdf)
    A Hybrid Convolutional Approach for Parking Availability Prediction.
    In 2019 International Joint Conference on Neural Networks (IJCNN) (pp. 1-8). IEEE.
  • Rashed, A., Grabocka, J. and Schmidt-Thieme, L., 2019, February. (pdf)
    Weighted Personalized Factorizations for Network Classification with Approximated Relation Weights.
    In International Conference on Agents and Artificial Intelligence (pp. 100-117). Springer, Cham.
  • Grabocka, J., Scholz, R. and Schmidt-Thieme, L., 2019. (arXiv)
    Learning surrogate losses. 
    arXiv preprint arXiv:1905.10108.
  • Jomaa, H.S., Schmidt-Thieme, L. and Grabocka, J., 2019. (arXiv)
    Dataset2vec: Learning dataset meta-features. 
    arXiv preprint arXiv:1905.11063.
  • Brinkmeyer, L., Drumond, R.R., Scholz, R., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    Chameleon: Learning model initializations across tasks with different schemas. 
    arXiv preprint arXiv:1909.13576.
  • Jomaa, H.S., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    Hyp-rl: Hyperparameter optimization by reinforcement learning. 
    arXiv preprint arXiv:1906.11527.
  • Jawed, S., Boumaiza, E., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    Data-driven vehicle trajectory forecasting. 
    arXiv preprint arXiv:1902.05400.
  • Rashed, A., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    Multi-label network classification via weighted personalized factorizations. 
    arXiv preprint arXiv:1902.09294.
  • Rego Drumond, R., Brinkmeyer, L., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    HIDRA: Head Initialization across Dynamic targets for Robust Architectures.
    arXiv, pp.arXiv-1910.
  • Jomaa, H.S., Grabocka, J. and Schmidt-Thieme, L., 2019. (arXiv)
    In Hindsight: A Smooth Reward for Steady Exploration. 
    arXiv preprint arXiv:1906.09781.
  • Falkner, J., Grabocka, J. and Schmidt-Thieme, L., 2019. (pdf)
    Atomic Compression Networks.
  • Feurer, M., van Rijn, J.N., Kadra, A., Gijsbers, P., Mallik, N., Ravi, S., Müller, A., Vanschoren, J. and Hutter, F., 2019. (arXiv)
    OpenML-Python: an extensible Python API for OpenML. 
    arXiv preprint arXiv:1911.02490.

2018

  • Grabocka, J. and Schmidt-Thieme, L., 2018. Neuralwarp. (arXiv)
    Time-series similarity with warping networks. 
    arXiv preprint arXiv:1812.08306.

2017

  • Nguyen, H.T., Wistuba, M., Grabocka, J., Drumond, L.R. and Schmidt-Thieme, L., 2017, May. (pdf)
    Personalized deep learning for tag recommendation.
    In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 186-197). Springer, Cham.
  • Raychaudhuri, D.S., Grabocka, J. and Schmidt-Thieme, L., 2017. (arXiv)
    Channel masking for multivariate time series shapelets. 
    arXiv preprint arXiv:1711.00812.

2016

  • Shah, M., Grabocka, J., Schilling, N., Wistuba, M. and Schmidt-Thieme, L., 2016, March. (pdf)
    Learning DTW-shapelets for time-series classification.
    In Proceedings of the 3rd IKDD Conference on Data Science, 2016 (pp. 1-8).
  • Grabocka, J., Schilling, N. and Schmidt-Thieme, L., 2016. (pdf)
    Latent time-series motifs. 
    ACM Transactions on Knowledge Discovery from Data (TKDD), 11(1), pp.1-20.
  • Grabocka, J., Wistuba, M. and Schmidt-Thieme, L., 2016. (pdf)
    Fast classification of univariate and multivariate time series through shapelet discovery. 
    Knowledge and information systems, 49(2), pp.429-454.
  • Grabocka, J., 2016. (pdf)
    Invariant Features For Time-Series Classification 
    (Doctoral dissertation, Universität Hildesheim).

2015

  • Wistuba, M., Grabocka, J. and Schmidt-Thieme, L., 2015. (arXiv)
    Ultra-fast shapelets for time series classification. 
    arXiv preprint arXiv:1503.05018.
  • Grabocka, J. and Schmidt-Thieme, L., 2015. (pdf)
    Learning through non-linearly supervised dimensionality reduction.
    In Transactions on Large-Scale Data-and Knowledge-Centered Systems XVII (pp. 74-96). Springer, Berlin, Heidelberg.

2014

  • Grabocka, J., Dalkalitsis, A., Lois, A., Katsaros, E. and Schmidt-Thieme, L., 2014, October. (pdf)
    Realistic optimal policies for energy-efficient train driving.
    In 17th International IEEE Conference on Intelligent Transportation Systems (ITSC) (pp. 629-634). IEEE.
  • Grabocka, J., Schilling, N., Wistuba, M. and Schmidt-Thieme, L., 2014, August. (pdf)
    Learning time-series shapelets.
    In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 392-401).
  • Grabocka, J., Bedalli, E. and Schmidt-Thieme, L., 2014, May. (pdf)
    Supervised nonlinear factorizations excel in semi-supervised regression.
    In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 188-199). Springer, Cham.
  • Grabocka, J., Wistuba, M. and Schmidt-Thieme, L., 2014. (pdf)
    Scalable classification of repetitive time series through frequencies of local polynomials. 
    IEEE Transactions on Knowledge and Data Engineering, 27(6), pp.1683-1695.
  • Grabocka, J. and Schmidt-Thieme, L., 2014. (pdf)
    Invariant time-series factorization.
    Data mining and knowledge discovery, 28(5-6), pp.1455-1479.

2013

  • Grabocka, J., Drumond, L. and Schmidt-Thieme, L., 2013, August. (pdf)
    Supervised dimensionality reduction via nonlinear target estimation.
    In International Conference on Data Warehousing and Knowledge Discovery (pp. 172-183). Springer, Berlin, Heidelberg.

2012

  • Grabocka, J., Nanopoulos, A. and Schmidt-Thieme, L., 2012, September. (pdf)
    Invariant time-series classification.
    In Joint European Conference on Machine Learning and Knowledge Discovery in Databases (pp. 725-740). Springer, Berlin, Heidelberg.
  • Grabocka, J., Bedalli, E. and Schmidt-Thieme, L., 2012, September. (pdf)
    Efficient classification of long time-series.
    In International Conference on ICT Innovations (pp. 47-57). Springer, Berlin, Heidelberg.
  • Grabocka, J., Nanopoulos, A. and Schmidt-Thieme, L., 2012, July. (pdf) 
    Classification of sparse time series via supervised matrix factorization. 
    In Twenty-Sixth AAAI Conference on Artificial Intelligence.