Entrance hall of the ICMS

Distributed machine learning and optimization

Nov 25, 2015 - Nov 27, 2015

ICMS, 15 South College Street

Organisers

Name Institution
Czumaj, Artur University of Warwick
Diakonikolas, Ilias University of Edinburgh
Girolami, Mark University of Warwick
Hauser, Raphael University of Oxford
Richtarik, Peter University of Edinburgh
Shawe-Taylor, John University College London


Modern data sets are often too large to be stored on a single machine and are typically distributed among many servers. Surprisingly little is known about how to properly approach massive-scale data analysis tasks in distributed environments. The current practice is to either rely on heuristics with no theoretical guarantees or to apply classical methods which are known to work well in medium-scale. However, heuristics often exhibit unpredictable behaviour, while classical algorithms are often impractical or impossible to apply in a massive-scale domain. Successful approach to “making sense from data” problems at massive scales requires the merging of ideas from multiple scientific disciplines. The “Theoretical Foundations of Big Data Analysis” programme held at the Simons Institute for the Theory of Computing in 2013 (Aug–Dec)—where some of the organizers participated as long-term invited visiting scientists—identified three pillars of big data analysis: computer science, optimization and statistics. Our workshop brings together researchers in these three disciplines as well as industry practitioners, with the aim to review the state of the art, identify key theoretical and practical challenges and outline a programme of future work to be conducted by the Alan Turing Institute in this key area of big data analytics.

Key scientific question to be answered and key topics to be addressed:

  • Review state-of-the-art algorithms for distributed optimization

  • Review state-of-the-art distributed algorithms for key machine learning tasks
  • Identify key theoretical challenges in the field, important open problems and most promising avenues for future research and progress
  • Identify most burning issues faced by industry and the most promising solutions academia can offer


Topics: algorithms, modelling, applications, systems, complexity, scalability, big data

Key sectors involved and impacted: information technology, data science, energy, e-commerce, healthcare, biology, computer vision, finance, education, government

This workshop is one of a number of scientific scoping workshops will be held, at the British Library and at other locations across the UK by the Alan Turing Institute. These workshops, which were approved via a competitive process, will map out the national and international data science landscape, focussing on areas core to the Institute’s mission, including computer science and informatics, the mathematical sciences, social science and ethics. They will also serve as an instrument for the development of a number of coherent research programmes.

 

Arrangements

Participation

INFORMATION FOR INVITED PARTICIPANTS

  • Invited participants will have received an emailed invitation from the Scientific Organisers and an "invitation to register" from ICMS.  The ICMS invitation contains a link to your individual registration form.    

INFORMATION FOR PUBLIC APPLICANTS

  • The public application form is available here.   Please note that we are unable to offer financial assistance.  However, it is intended that the catering (lunch and refreshments each day) will be provided and NO registration fee will be charged.  Some suggestions for accommodation in Edinburgh are given in the "Accommodation" section.  

 

Venue and Talks
The workshop will be held at 15 South College Street, Edinburgh.  Talks will be held in the Newhaven Lecture Theatre.    The Lecture Theatre is equipped with a data projector, computer, visualiser (the new generation of overhead projectors) and two blackboards.  The projector and one board may be used simultaneously.  It is best to bring your presentation on a memory stick to use in our ICMS computer.   Alternatively, it is possible for you to use your own laptop with our data projector, but please be aware that you may have to alter your laptop resolutions/settings. 

 

UK Visas
If you are travelling from overseas you may require an entry visa. A European visa does not guarantee entry to the UK. Please use this link to the UK Visas site to find out if you need a visa and if so how to apply for one.

 

Travel
Information about travel to the UK and Edinburgh is available here.  Please note that it is your responsibility to have adequate travel insurance to cover medical and other emergencies that may occur on your trip.

A taxi directly from the airport will cost approximately 20.00 GBP to the city centre for a one-way journey.     There is also a bus service direct from the airport to the city centre which will cost 4.50 single or 7.50 GBP return - the Airlink 100.  This is a frequent service (every 10 minutes during peak times) and will bring you close to Waverley Railway Station and the workshop venue. 

Lothian buses charge £1.50 for a single, £4.00 for a day ticket. Please note that the exact fare is required and no change is given.

If travelling by train, please note that Edinburgh has several railway stations; Waverley Railway Station being the main station and closest to the workshop venue at 15 South College Street. If you alight at Edinburgh Waverley, the meeting venue is an easy 10 minute walk over North and South Bridge.  The other railway stations are Haymarket and Edinburgh Park but please note that these stations are at the West End of the city centre.

 

Accommodation

  • ICMS will organise accommodation for Invited Participants (if requested on your registration form).
  • Public Applicants are asked to arrange and pay for your own accommodation (once your application has been accepted).   Some suggestions are below:

    Masson House, Pollock Halls of Residence, The University of Edinburgh, 18 Holyrood Park Road, Edinburgh EH16 5AY  (approx 70.00 GBP per night)
    0800 028 7118 (UK only) - +44 (0)131 651 2189 - http://www.book.accom.ed.ac.uk/
    University of Edinburgh accommodation. Pollock Halls is about a 20-25 minute walk from ICMS.

    Jurys Inn Edinburgh, 43 Jeffrey Street, Edinburgh EH1 1DH (approx 75.00-90.00 GBP per night)
    +44 (0)131 200 3300 - www.jurys-edinburgh-hotels.com
    Only a 5 minute walk from ICMS.

    Edinburgh City Hotel, 79 Lauriston Place, Edinburgh, EH3 9HZ  (approx 70-80 GBP per night)
    +44 (0)131 622 7979 - http://www.edinburghcityhotel.com/

    St Christophers Hostel, 9-13 Market Street, Edinburgh EH1 1DE  (approx 25.00 or less depending on offers available)
    +44 (0)20 7407 1856 -  www.st-christophers.co.uk - bookings@st-christophers.co.uk
    This is Youth Hostel accommodation.

Further details and many more suggestions for accommodation can be found here 

 

Catering
Refreshments throughout the event and lunch each day will be provided.  

 

Wireless Access
Access to wifi via Eduroam is available throughout the building. If you are not registered with Eduroam you will be given instructions and a code for accessing the wireless network.  For those without laptops, there will also be a couple of computers available for you to check your emails.  

 

Registration Fee
There is no registration fee payable for this event.

 

Programme

Please continue to check the Programme below as it may be subject to alteration.


Wednesday 25 November

09:00-09:45

Registration and coffee in the Chapterhouse, Level 1

09:45-10:00

Opening remarks (Peter Richtarik)

10:00-10:35

David Woodruff (IBM Research - Almaden)
A story of principal component analysis in the distributed model

10:35-11:10

Jakub Konecny (University of Edinburgh)   
Federated optimization: distributed optimization beyond the datacenter              

11:10-11:30

Coffee in the Chapterhouse                 

11:30-12:05

Yee Whye Teh (University of Oxford)
On distributed MCMC and variational machine learning

12:05-12:30

Summary

12:30-13:30

Lunch provided in the Chapterhouse

13:30-14:05

Raphael Hauser (University of Oxford)
Computing the leading part SVD of large dense matrices

14:05-14:40

Michal Kocvara (University of Birmingham)
On multigrid methods in convex optimization

14:40-15:15

Yves Wiaux (Heriot-Watt University/University of Edinburgh)
Astronomical imaging - in every sense of the word: scalable optimization algorithms in radio-interferometry

15:15-15:45

Coffee in the Chapterhouse

15:45-17:00

Parallel discussions

17:00-17:30

Summary

17:30-18:30

Informal wine reception in the Chapterhouse

 


Thursday 26 November

09:30-10:05

Gregory Valiant (Stanford University)
Memory, communication, and statistical queries

10:05-10:40

Robert Gower (University of Edinburgh)
Randomized iterative methods for linear systems

10:40-11:10

Coffee in the Chapterhouse               

11:10-11:45

Krzysztof Onak (IBM Thomas J Watson Research Center)
Communication complexity of learning discrete distributions              

11:45-12:20

Coralia Cartis (University of Oxford)
Optimization algorithms with only occasionally accurate models

12:20-13:30

Lunch provided in the Chapterhouse

13:30-14:05

Alina Ene (University of Warwick)
A new framework for distributed submodular maximization

14:05-14:40

Christian Sohler (University of Dortmund)
Testing cluster structure of graphs

14:40-15:15

Coffee in the Chapterhouse

15:15-17:00

Parallel discussions

17:00-17:30

Summary

 

Friday 27 November

09:30-10:05

Cedric Archambeau (Amazon)
Incremental variational inference: a distributed probabilistic inference framework

10:05-10:40

Amos Storkey (University of Edinburgh)
Mechanism design for distributing machine learning

10:40-11:10

Coffee in the Chapterhouse               

11:10-11:45

Artur Czumaj (University of Warwick)
Testing directed networks            

11:45-12:20

He Sun (University of Bristol)
Partitioning well-clustered graphs: spectral clustering works!

12:20-13:30

Lunch provided in the Chapterhouse

13:30-15:00

Discussions

15:00-15:30

Coffee in the Chapterhouse

15:30-17:00

Parallel discussions

17:00-17:30

Summary

 

Participants

Name Institution
Abdullah, Mohammed Huawei Technologies Co Ltd
Archambeau, Cedric Amazon
Cartis, Coralia University of Oxford
Chang, Ya-Ting University of Edinburgh
Chouvardas, Symeon Huawei Technologies Co Ltd
Csiba, Dominik University of Edinburgh
Czumaj, Artur University of Warwick
Dabbech, Arwa Heriot-Watt University
Damoulas, Theo University of Warwick
Diakonikolas, Ilias University of Edinburgh
Ene, Alina University of Warwick
Etessami, Kousha University of Edinburgh
Gazzola, Silvia Heriot-Watt University
Gower, Robert University of Edinburgh
Grothey, Andreas University of Edinburgh
Hall, Julian University of Edinburgh
Hauser, Raphael University of Oxford
Kashyap, Sidharth University of Edinburgh
Kocvara, Michal University of Birmingham
Konecny, Jakub University of Edinburgh
Lisowska, Aneta Toshiba Medical Visualization Systems Europe
Loizou, Nicolas University of Edinburgh
McKinnon, Kenneth University of Edinburgh
Mei, Zhaoxun Heriot-Watt University
Mokry, Jozef University of Edinburgh
Nikishkin, Vladimir University of Edinburgh
Onak, Krzysztof IBM Thomas J Watson Research Center
Onose, Alexandru Heriot-Watt University
Pavlakou, Theo The University of Edinburgh
Pavlou, Chrystalla University of Edinburgh
Repetti, Audrey Heriot-Watt University
Richtarik, Peter University of Edinburgh
Sohler, Christian Technische Universit├Ąt Dortmund
Stewart, Alistair University of Edinburgh
Storkey, Amos University of Edinburgh
Sun, He University of Bristol
Tan, Shuren University of Edinburgh
Teh, Yee Whye University of Oxford
Valiant, Gregory Stanford University
Wiaux, Yves Heriot-Watt University/University of Edinburgh
Wirhaspati, Novia Listiyani University of Edinburgh
Woodruff, David IBM Research - Almaden
Zhu, Zhanxing University of Edinburgh