Current research
Machine learning of Big Data, Internet of Things, Swarm robotics, and 3D-cameras.
I am especially excited about reducing the gap between theoretical and practical algorithms, using my experience in
the industry and academy.
Main technique
Core-sets: Semantic compression of data sets into small sets that provably approximate
the original data for a given problem. Using merge-reduce (e.g. Spark) the small sets
can then be used for solving hard machine learning problems in parallel (on the cloud/network),
real-time and on Big streaming data.
Introduction to Coresets
Videos from the Machine Learning Summer School 2014 at CMU
Lecture 1:
VIDEO
Lectures 2 and 3:
VIDEO
Lecture 4:
VIDEO
Lecture 5:
VIDEO
Publications and Slides
Fast and Accurate Least-Mean-Squares Solvers
[slides]
with Alaa Maalouf and Ibrahim Jubran.
[oral presentation]
[Short Interview]
Proc. 33rd Conference on Neural Information Processing Systems (NIPS) 2019 (Oral + Outstanding Paper Honorable Mention)
Introduction to Coresets: Accurate Coresets
with Ibrahim Jubran and Alaa Maalouf
(under review)
Tight Sensitivity Bounds For Smaller Coresets
with Alaa Maalouf and Adiel Statman
Provable Filter Pruning for Efficient Neural Networks
Lucas Liebenwein, Cenk Baykal, Harry Lang, Daniela Rus.
Data-Independent Neural Pruning via Coresets
with Ben Mussay, Margarita Osadchy, Vladimir Braverman, Samson Zhou.
Secure Data Retrieval on the Cloud: Homomorphic Encryption meets Coresets
with Adi Akavia, Hayim Shaul
Transactions on Cryptographic Hardware and Embedded Systems, Accepted for Publication
Position Estimation of Multiple Robots: Provable, Practical Approximation Algorithm
With Ariel Hutterer, Danial, Jeryes.
IEEE Robotics and Automation Letters (RA-L), Accepted for Publication
Measuring Privacy in the Always-On Era
With Eldar Haber
Cardozo Arts & Entertainment Law Journal, Accepted for Publication
Real-Time EEG Classification via Coresets for BCI Applications
with Eitan Netzer, Alex Frid.
Minimizing Sum of Non-Convex but Piecewise log-Lipschitz Functions using Coresets
with Ibrahim Jubran.
[Video] ,
[Presentation]
Secure Search via Sketching for Homomorphic Encryption
with Adi Akavia, Hayim Shaul
25th ACM Conference on Computer and Communications Security 2018.
Coresets For Monotonic Functions with Applications to Deep Learning
with Elad Tolochinsky, Submitted.
Quadcopter Tracks Quadcopter via Real Time Shape Fitting
with Dror Epstein.
[Video]
IEEE Robotics and Automation Letters (RA-L) 2017
Secure Search on the Cloud via Coresets and Sketches
with Adi Akavia, Hayim Shaul
Training Mixture Models at Scale via Coresets (Fuller version)
with Mario Lucic, Matthew Faulkner, Andreas Krause
Journal of Machine Learning (JMLR) 2017
Coresets for Vector Summarization with Applications to Network Graphs
with Sedat Ozer, and Daniela Rus
International Conference on Machine Learning (ICML) 2017
Coresets for Differentially Private K-Means Clustering and Applications to Privacy in Mobile Sensor Networks
with Ruihao Zhu, Chongyuan Xiang and Daniela Rus
ACM/IEEE Conf. on Information Processing in Sensor Networks (IPSN) 2017, to appear
New Frameworks for Offline and Streaming Coreset Constructions
with Vladimir Braverman, Harry Lang
Dimensionality Reduction of Massive Sparse Datasets Using Coresets
with Mikhali Volkov, Daniela Rus, (supplementary material)
Proc. 29th Conference on Neural Information Processing Systems (NIPS) 2016
k -Means for Streaming and Distributed Big Sparse Data
with Artem Barger
Proceedings of the 2016 SIAM International Conference on Data Mining (SDM) 2016
Low-cost and Faster Tracking Systems Using Core-sets for Pose-Estimation
with Soliman Nasser and Ibrahim Jubran [Video]
iDiary: From GPS Signals to a Text-Searchable Diary
with Andrew Sugaya, Cynthia Sung, and Daniela Rus
ACM Transactions on Sensor Networks, Volume I, Issue 4 2015
Coresets for Visual Summarization with Applications to Loop Closure
with G. Rossman, Dan Feldman, Mikhail V. Volkov and D. Rus
IEEE International Conference on Robotics and Automation (ICRA) 2015
Fleye on the Car: Big data Meets the Internet of Things
with Soliman Nasser, Andrew Barry, Marek Doniec, Guy Peled, Guy Rosman, Daniela Rus,
and Mikhail Volkov.
ACM/IEEE Conf. on Information Processing in Sensor Networks (IPSN) 2015
More Constraints, Smaller Coresets: Constrained Matrix Approximation of Sparse Big Data
with Tamir Tassa
21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) 2015
Coresets for k -segmentation of streaming data
with G. Rosman, M. Volkov, J.W. Fisher III, and D. Rus
Proc. 27th Conference on Neural Information Processing Systems (NIPS) 2014
Smallest enclosing ball for probabilistic data
with Alexander Munteanu, and Christian Sohler
The 30th Annual Symposium on Computational Geometry (SoCG) 2014
Visual Precis Generation using Coresets
with Rohan Paul, Daniela Rus and Paul Newman.
IEEE International Conference on Robotics and Automation (ICRA) 2014
Big Data for Robots: Online HMM Coresets for Sensor Streams
with Cathy Wu, Brian Julian, Cynthia Sung, and Daniela Rus
IEEE International Conference on Robotics and Automation (ICRA) 2013
K-Robots Clustering of Moving Sensors Using Coresets
, [Slides]
with Stephanie Gil and Daniela Rus
IEEE International Conference on Robotics and Automation (ICRA) 2013
iDiary: From GPS Signals to a Text-Searchable iDiary
with Andrew Sugaya, Cynthia Sung, and Daniela Rus
The 11th ACM Conference on Embedded Networked Sensor System (SenSys) 2013
Learning Big (Image) Data via Coresets for Dictionaries
with Micha Feigin, and Nir Sochen
The International Biomedical and Astronomical Signal Processing (BASP) Frontiers workshop, 2013, selected to
The Journal of Mathematical Imaging and Vision (JMIV)
The Single Pixel GIS: Learning Big Data Signals from Tiny Coresets
with Cynthia Sung and Daniela Rus
Proc. 20th ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL GIS) 2012
Turning Big Data into Tiny Data:
Constant-size Coresets for k -means, PCA and Projective Clustering
with Melanie Schmidt and Christian Sohler
Proc. 24th Annu. ACM Symp. on Discrete Algorithms (SODA) 2013,
and in the 4th Workshop on Massive Data Algorithms (MASSIVE) 2012
Trajectory Clustering for Motion Prediction
with Cynthia Sung and Daniela Rus
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2012
Communication Coverage for Independently Moving Robots
with Stephanie Gil and Daniela Rus
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2012
An Effective Coreset Compression Algorithm for Large Scale Sensor Networks,
with Andrew Sugaya and Daniela Rus
Proc. 11th ACM/IEEE Conf. on Information Processing in Sensor Networks (IPSN) 2012
Data Reduction for Weighted and Outlier-resistant Clustering,
[Slides]
with Leonard J. Schulman.
Proc. 23th Annu. ACM Symp. on Discrete Algorithms (SODA) 2012
Scalable Training of Mixture Models via Coresets
[Slides]
with Andreas Krause and Matthew Faulkner.
Proc. 25th Conference on Neural Information Processing Systems (NIPS) 2011
A Unified Framework for Approximating and Clustering Data
with Michael Langberg
Proc. 43st Annu. ACM Symposium on Theory of Computing (STOC) 2011
[Fuller Version]
From High Definition Image to Low Space Optimization
with Micha Feigin and Nir Sochen
Scale Space and Variational Methods in Computer Vision (SSVM) 2011
Coresets and Sketches for High Dimensional Subspace Approximation Problems
with Morteza Monemizadeh, Christian Sohler and David Woodruf,
Proc. 21th Annu. ACM Symp. on Discrete Algorithms (SODA) 2010
Private Coresets,
[Slides]
with Amos Fiat, Haim Kaplan and Kobbi Nissim
Proc. 41st Annu. ACM Symposium on Theory of Computing (STOC) 2009
A PTAS for k-Means Clustering Based on Weak Coresets
with Morteza Monemizadeh and Christian Sohler
Proc. 23th Annu. ACM Symposium on Computational Geometry (SoCG) 2007
Bi-criteria Linear-time Approximations for Generalized k-Mean/Median/Center
with Amos Fiat, Danny Segev and Micha Sharir
Proc. 23th Annu. ACM Symposium on Computational Geometry (SoCG) 2007
Coresets for Weighted Facilities and Their Applications
[Slides]
with Amos Fiat and Micha Sharir
Proc. 47th Annu. IEEE Symposium on Foundations of Computer Science (FOCS) 2006
Resume
Dan Feldman is a faculty member and the head of the new Robotics & Big Data Labs in the University
of Haifa, after returning from a 3 years post-doc at the robotics lab of MIT. During his PhD in the
University of Tel-Aviv he developed data reduction techniques known as core-sets,
based on computational geometry. Since his post-docs at Caltech and MIT, Dan's coresets are applied
for main problems in Machine Learning, Big Data, computer vision, EEG and robotics. His group in Haifa
continues to design and implement core-sets with provable guarantees for such real-time systems.