[Theory] Reminder: 5/13 Thesis Defense: Blake Woodworth, TTIC

[Mary Marre]

*Thesis Defense: Blake Woodworth, TTIC*

*When:*      Thursday*,* May 13th at 9:00 am CT

*Where:*    * Join virtually here
<https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09>*

*Who: *       Blake Woodworth, TTIC

*Title:        *The Minimax Complexity of Distributed Optimization

*Abstract: *With the increasingly large scale of optimization problems,
particularly those arising from training huge machine learning models using
massive datasets, it is important to leverage parallelism to tractably
solve these problems. This thesis focuses on understanding optimal
algorithms for and the fundamental limits of distributed optimization. I
will describe the "graph oracle model" framework, an extension of the
classic oracle model to the distributed setting, and apply this framework
to study several distributed optimization settings of interest. I will
focus on the intermittent communication setting---where several machines
optimize in parallel but with limited communication between them---where I
will characterize the optimal error and optimal algorithms. Finally, I will
describe several possible avenues for circumventing these lower bounds in
order to be "better than optimal."

*Thesis Advisor:* Nathan Srebro <nati at ttic.edu>

******************************************************************************************************

Zoom link for the virtual presentation.
https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09
Meeting ID: 994 8586 3831
Passcode: 091555


Mary C. Marre
Faculty Administrative Support
*Toyota Technological Institute*
*6045 S. Kenwood Avenue*
*Chicago, IL  60637*
*mmarre at ttic.edu <mmarre at ttic.edu>*


On Wed, May 12, 2021 at 5:30 PM Mary Marre <mmarre at ttic.edu> wrote:

> *Thesis Defense: Blake Woodworth, TTIC*
>
> *When:*      Thursday*,* May 13th at 9:00 am CT
>
> *Where:*    * Join virtually here
> <https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09>*
>
> *Who: *       Blake Woodworth, TTIC
>
> *Title:        *The Minimax Complexity of Distributed Optimization
>
> *Abstract: *With the increasingly large scale of optimization problems,
> particularly those arising from training huge machine learning models using
> massive datasets, it is important to leverage parallelism to tractably
> solve these problems. This thesis focuses on understanding optimal
> algorithms for and the fundamental limits of distributed optimization. I
> will describe the "graph oracle model" framework, an extension of the
> classic oracle model to the distributed setting, and apply this framework
> to study several distributed optimization settings of interest. I will
> focus on the intermittent communication setting---where several machines
> optimize in parallel but with limited communication between them---where I
> will characterize the optimal error and optimal algorithms. Finally, I will
> describe several possible avenues for circumventing these lower bounds in
> order to be "better than optimal."
>
> *Thesis Advisor:* Nathan Srebro <nati at ttic.edu>
>
>
> ******************************************************************************************************
>
> Zoom link for the virtual presentation.
> https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09
> Meeting ID: 994 8586 3831
> Passcode: 091555
>
>
>
>
> Mary C. Marre
> Faculty Administrative Support
> *Toyota Technological Institute*
> *6045 S. Kenwood Avenue*
> *Chicago, IL  60637*
> *mmarre at ttic.edu <mmarre at ttic.edu>*
>
>
> On Tue, May 11, 2021 at 9:32 AM Mary Marre <mmarre at ttic.edu> wrote:
>
>> *Thesis Defense: Blake Woodworth, TTIC*
>>
>> *When:*      Thursday*,* May 13th at 9:00 am CT
>>
>> *Where:*    * Join virtually here
>> <https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09>*
>>
>> *Who: *       Blake Woodworth, TTIC
>>
>> *Title:        *The Minimax Complexity of Distributed Optimization
>>
>> *Abstract: *With the increasingly large scale of optimization problems,
>> particularly those arising from training huge machine learning models using
>> massive datasets, it is important to leverage parallelism to tractably
>> solve these problems. This thesis focuses on understanding optimal
>> algorithms for and the fundamental limits of distributed optimization. I
>> will describe the "graph oracle model" framework, an extension of the
>> classic oracle model to the distributed setting, and apply this framework
>> to study several distributed optimization settings of interest. I will
>> focus on the intermittent communication setting---where several machines
>> optimize in parallel but with limited communication between them---where I
>> will characterize the optimal error and optimal algorithms. Finally, I will
>> describe several possible avenues for circumventing these lower bounds in
>> order to be "better than optimal."
>>
>> *Thesis Advisor:* Nathan Srebro <nati at ttic.edu>
>>
>>
>> ******************************************************************************************************
>>
>> Zoom link for the virtual presentation.
>>
>> https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09
>> Meeting ID: 994 8586 3831
>> Passcode: 091555
>>
>>
>>
>>
>>
>> Mary C. Marre
>> Faculty Administrative Support
>> *Toyota Technological Institute*
>> *6045 S. Kenwood Avenue*
>> *Room 517*
>> *Chicago, IL  60637*
>> *p:(773) 834-1757*
>> *f: (773) 357-6970*
>> *mmarre at ttic.edu <mmarre at ttic.edu>*
>>
>>
>> On Fri, Apr 30, 2021 at 10:10 AM Mary Marre <mmarre at ttic.edu> wrote:
>>
>>> *Thesis Defense: Blake Woodworth, TTIC*
>>>
>>> *When:*      Thursday*,* May 13th at 9:00 am CT
>>>
>>> *Where:*    * Join virtually here
>>> <https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09>*
>>>
>>> *Who: *       Blake Woodworth, TTIC
>>>
>>> *Title:        *The Minimax Complexity of Distributed Optimization
>>>
>>> *Abstract: *With the increasingly large scale of optimization problems,
>>> particularly those arising from training huge machine learning models using
>>> massive datasets, it is important to leverage parallelism to tractably
>>> solve these problems. This thesis focuses on understanding optimal
>>> algorithms for and the fundamental limits of distributed optimization. I
>>> will describe the "graph oracle model" framework, an extension of the
>>> classic oracle model to the distributed setting, and apply this framework
>>> to study several distributed optimization settings of interest. I will
>>> focus on the intermittent communication setting---where several machines
>>> optimize in parallel but with limited communication between them---where I
>>> will characterize the optimal error and optimal algorithms. Finally, I will
>>> describe several possible avenues for circumventing these lower bounds in
>>> order to be "better than optimal."
>>>
>>> *Thesis Advisor:* Nathan Srebro <nati at ttic.edu>
>>>
>>>
>>> ******************************************************************************************************
>>>
>>> Zoom link for the virtual presentation.
>>>
>>> https://uchicago.zoom.us/j/99485863831?pwd=U2pDL2hpTW5URnppbFJBb2dxYnNFdz09
>>> Meeting ID: 994 8586 3831
>>> Passcode: 091555
>>>
>>>
>>>
>>> Mary C. Marre
>>> Faculty Administrative Support
>>> *Toyota Technological Institute*
>>> *6045 S. Kenwood Avenue*
>>> *Chicago, IL  60637*
>>> *mmarre at ttic.edu <mmarre at ttic.edu>*
>>>
>>>
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/theory/attachments/20210513/16b2597c/attachment-0001.html>