From sekhar.tatikonda at yale.edu  Mon Feb  4 19:31:41 2019
From: sekhar.tatikonda at yale.edu (Sekhar Tatikonda)
Date: Mon, 4 Feb 2019 19:31:41 -0500
Subject: [YPNG] YPNG Friday 8 February 2019
In-Reply-To: <5DDC39FA-DF84-41E8-BD97-DE186CC3E3E0@yale.edu>
References: <5DDC39FA-DF84-41E8-BD97-DE186CC3E3E0@yale.edu>
Message-ID: <CANaQW2AoSGSVDd3h_A5pgV71NM2i3S+U+1XR0eYKcNMKNmjO9g@mail.gmail.com>

Hi Folks,

This Friday, Feb 8, and next Friday, Feb 15, students will recap the events
at the MIT workshop on non-convex optimization and deep learning. There
will
be 5-6 presentations, each lasting about 10 minutes and giving a high-level
summary of one of the workshop talks.

The broad topics of the workshop were generalization, expressive power, and
optimization. This Friday, we'll have the presentations on generalization
and
expressive power and a couple on optimization. Next Friday it will be all
optimization.

See you Friday at 11am in the Stat's classroom.

Regards,
sekhar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.yale.edu/pipermail/ypng/attachments/20190204/bb6c21c9/attachment.html>

From sekhar.tatikonda at yale.edu  Tue Feb 12 09:30:33 2019
From: sekhar.tatikonda at yale.edu (Sekhar Tatikonda)
Date: Tue, 12 Feb 2019 09:30:33 -0500
Subject: [YPNG] YPNG Friday 15 February 2019
In-Reply-To: <CANaQW2AoSGSVDd3h_A5pgV71NM2i3S+U+1XR0eYKcNMKNmjO9g@mail.gmail.com>
References: <5DDC39FA-DF84-41E8-BD97-DE186CC3E3E0@yale.edu>
 <CANaQW2AoSGSVDd3h_A5pgV71NM2i3S+U+1XR0eYKcNMKNmjO9g@mail.gmail.com>
Message-ID: <CANaQW2DFV3wv6pLUNJ7K5ncnHrvrPazfqXePpjEUWhGmMz1W2g@mail.gmail.com>

Hi Everyone,
Round 2 of the student presentations continue this Friday at 11am in the
Stat's classroom.
(See description below.)

Regards,
sekhar




---------- Forwarded message ---------
From: Sekhar Tatikonda <sekhar.tatikonda at yale.edu>
Date: Mon, Feb 4, 2019 at 7:31 PM
Subject: YPNG Friday 8 February 2019
To: <ypng at mailman.yale.edu>




Hi Folks,

This Friday, Feb 8, and next Friday, Feb 15, students will recap the events
at the MIT workshop on non-convex optimization and deep learning. There
will
be 5-6 presentations, each lasting about 10 minutes and giving a high-level
summary of one of the workshop talks.

The broad topics of the workshop were generalization, expressive power, and
optimization. This Friday, we'll have the presentations on generalization
and
expressive power and a couple on optimization. Next Friday it will be all
optimization.

See you Friday at 11am in the Stat's classroom.

Regards,
sekhar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.yale.edu/pipermail/ypng/attachments/20190212/50ca967a/attachment-0001.html>

From sekhar.tatikonda at yale.edu  Mon Feb 18 11:37:08 2019
From: sekhar.tatikonda at yale.edu (Sekhar Tatikonda)
Date: Mon, 18 Feb 2019 11:37:08 -0500
Subject: [YPNG] YPNG Friday 22 February 2019
In-Reply-To: <CANaQW2DFV3wv6pLUNJ7K5ncnHrvrPazfqXePpjEUWhGmMz1W2g@mail.gmail.com>
References: <5DDC39FA-DF84-41E8-BD97-DE186CC3E3E0@yale.edu>
 <CANaQW2AoSGSVDd3h_A5pgV71NM2i3S+U+1XR0eYKcNMKNmjO9g@mail.gmail.com>
 <CANaQW2DFV3wv6pLUNJ7K5ncnHrvrPazfqXePpjEUWhGmMz1W2g@mail.gmail.com>
Message-ID: <CANaQW2BxH7LrR9bfEV=sfqxqPF96jqPyVF0DeEtiTmyvYpgZWw@mail.gmail.com>

Hi Folks,
Our third (and final) installment of the student presentations continues
this Friday at 11am in the Stat's classroom.  See description below.

Regards,
sekhar





---------- Forwarded message ---------
From: Sekhar Tatikonda <sekhar.tatikonda at yale.edu>
Date: Tue, Feb 12, 2019 at 9:30 AM
Subject: YPNG Friday 15 February 2019
To: <ypng at mailman.yale.edu>




Hi Everyone,
Round 2 of the student presentations continue this Friday at 11am in the
Stat's classroom.
(See description below.)

Regards,
sekhar




---------- Forwarded message ---------
From: Sekhar Tatikonda <sekhar.tatikonda at yale.edu>
Date: Mon, Feb 4, 2019 at 7:31 PM
Subject: YPNG Friday 8 February 2019
To: <ypng at mailman.yale.edu>




Hi Folks,

This Friday, Feb 8, and next Friday, Feb 15, students will recap the events
at the MIT workshop on non-convex optimization and deep learning. There
will
be 5-6 presentations, each lasting about 10 minutes and giving a high-level
summary of one of the workshop talks.

The broad topics of the workshop were generalization, expressive power, and
optimization. This Friday, we'll have the presentations on generalization
and
expressive power and a couple on optimization. Next Friday it will be all
optimization.

See you Friday at 11am in the Stat's classroom.

Regards,
sekhar
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.yale.edu/pipermail/ypng/attachments/20190218/0f957d9c/attachment.html>

From david.pollard at yale.edu  Fri Feb 22 20:48:30 2019
From: david.pollard at yale.edu (David Pollard)
Date: Fri, 22 Feb 2019 20:48:30 -0500
Subject: [YPNG] Fwd: Simulated Tempering & Canonical Paths
References: <CAA_0ik14-zOdMS7YZWEWULU1K+-HUoG_PLmhe00VzCTvzEfePQ@mail.gmail.com>
Message-ID: <F9008435-19A7-4AF1-A288-48AC531D4AEC@yale.edu>

Folks,

Here is some of the work Dylan mentioned this morning.

David

> Begin forwarded message:
> 
> From: "Dylan O'Connell" <dylan.oconnell at yale.edu>
> Subject: Re: Simulated Tempering & Canonical Paths
> Date: February 22, 2019 at 2:33:11 PM EST
> To: David Pollard <david.pollard at yale.edu>
> 
> Sounds good to me. I'm sure the talk won't rely on it (in fact it seems to cover much of his work), but anyone curious about it might benefit from taking a glance before the presentation.
> 
> -Dylan
> 
> On Fri, Feb 22, 2019 at 1:55 PM David Pollard <david.pollard at yale.edu <mailto:david.pollard at yale.edu>> wrote:
> Thanks.  I put it on my dropbox at
> 
> https://www.dropbox.com/sh/y3xdpmltaxm35g1/AAA9xluS7vTZ2mCyPxooKlt5a?dl=0 <https://www.dropbox.com/sh/y3xdpmltaxm35g1/AAA9xluS7vTZ2mCyPxooKlt5a?dl=0>
> 
> Should I send the link to the YPNG mailing list?
> 
> David
> 
>> On Feb 22, 2019, at 1:45 PM, Dylan O'Connell <dylan.oconnell at yale.edu <mailto:dylan.oconnell at yale.edu>> wrote:
>> 
>> Dr. Pollard,
>> 
>> As promised, here <https://arxiv.org/abs/1812.00793> is a link to the arXiv for the presented paper. We have been continuing to discuss these results with the authors, and thus the proof details should not be considered gospel. Their latest update of the core Theorem 6.2 can be found here <https://www.dropbox.com/s/a0z779ewbcuuuhw/corrections-public.pdf?dl=0>, sent to us this morning. However, we have not confirmed this, nor does it include all relevant errata.
>> 
>> Despite the ongoing updates, all of the core ideas are contained in the arXiv paper (I could send a link to their poster or blog post if desired, but those summaries are likely to be out of date). Andrej Risteski will be speaking at the seminar on Monday, so while that is only a few days away, I'm sure there will be opportunity to discuss any questions about the work!
>> 
>> Sincerely,
>> Dylan
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.yale.edu/pipermail/ypng/attachments/20190222/eb7832bf/attachment.html>