We are the Distributed Learning Systems Lab at the Distributed Systems Section, EWI faculty, TU Delft. We working on a wide variety of research problems in distributed machine learning systems, and developing cutting edge AI technology for industry. Check our start-up spin-off, Generatrix.

Our lab is supported by TU Delft, Aegon, ABB Research, Swiss National Science Foundation (SNSF), and Dutch National Science Foundation (NWO).

News

[03/22] Our paper, “LegoDN" , is awareded with Runner-up Community Award at MobiCom

[02/22] Our paper, “Locality Sensitive Hash Aggregated Nonlinear Neighborhood MatrixFactorization for Online Sparse Big Data Analysis" , is accpted in ACM Transactions on Data Science "

[12/21] Our paper, “Lightweight and Accurate DNN-based Anomaly Detection at Edge" , is accpted in IEEE Transactions on Parallel and Distribution Computing

[12/21] Our paper, “EdgeTuner: Fast Scheduling Algorithm Tuning for Dynamic Edge-Cloud Workloads and Resources" , is accpted in IEEE INFOCOM22

[11/21] Our paper, “Locality Sensitive Hash Aggregated Nonlinear Neighborhood MatrixFactorization for Online Sparse Big Data Analysis" , is accpted in ACM Transactions on Data Science

[10/21] Lydia will serve TPC track co-chair of CCGRID22 .

[09/21] Our paper, CTAB-GAN: Effective Table Data Synthesizing , is accpted in ACML21

[09/21] Our paper, QActor: On-line Active Learning for Noisy Labeled Stream Data , is accpted in ACML21

[08/21] Our paper, LegoDNN: Block-grained Scaling of Deep Neural Networks for Mobile Vision , is accepted in MobiCom21

[08/21] Our paper, Federated Learning with Heterogeneity-Aware Probabilistic Synchronous Parallel on Edge , is accepted in IEEE Transactions on Service Computing

[08/21] Our paper, Disparity seeding , is accepted in CIKM21

[07/21] Lydia will serve TPC co-chair of BDCAT21 .

[08/21] Lydia will serve TPC Sigmetrics21 .

[07/21] The Generatrix team enters the semi-final of The European Social Innovation Competition !

[07/21] We graduate 13 bachelor students on broad topics of synthesizing comics, privacy preserving time-series synthesizing, and robust multi-label learning.

[07/21] Our paper on differential privacy for sparse tensor factorization is accepted in IEEE Transactions on Industrial Informatics

[05/21] Our PerCom'21 papre is invited for a special issuee of Pervasive and Mobile Computing Journal.

[05/21] Our paper, LABELNET: Recovering Noisy Labels, is accepted in IJCNN

[03/21] Our paper on differential private deep learning at edges is acceptedd in IEEE Transactions on Parallel and Distributed Computing

[02/21] Our paper, Enhancing Robustness of On-line Learning Models on Highly Noisy Data, is accepted in IEEE Transactions on Dependable and Secure Computing

[01/21] Our paper, Online Label Aggregation: A Variational Bayesian Approach, is accepted inWWW21

[01/21] Our two papers on Multiple Inference of Deep Models on Edge will appear in PerCom21

[12/20] We just received a NWO take-off grant for our project, "Tabular Data Synthesizer"

[12/20] Our paper on Practical Analysis of Replication-Based Systems will appear in INFOCOM2021

[11/20] Our paper, Courier: Real-Time Optimal Batch Size Prediction for Latency SLOs in BigDL, will appear in ICPE21

[11/20] Our paper on federated learnnig at heterogeneous platform will appear in AI/ML special issue of TPDS

[11/20] Our paper on stochastic optimization for parallel sparse Tucker decomposition will appear in AI/ML special issue of TPDS