On Thu, May 14, 2020 at 9:51 AM Michael Wong <fraggamuffin@gmail.com> wrote:

SG19 Machine Learning 2 hours

Michael Wong is inviting you to a scheduled Zoom meeting.


Michael Wong is inviting you to a scheduled Zoom meeting.

Topic: SG19 monthly Apr 2020-Oct 2020
Time: Apr 9, 2020 02:00 PM Eastern Time (US and Canada)  18:00 UTC
    Every month on the Second Thu, until Oct 8, 2020, 7 occurrence(s)
    Apr 9, 2020 02:00 PM 18:00 UTC
    May 14, 2020 02:00 PM  18:00 UTC
    Jun 11, 2020 02:00 PM  18:00 UTC
    Jul 9, 2020 02:00 PM  18:00 UTC
    Aug 13, 2020 02:00 PM  18:00 UTC
    Sep 10, 2020 02:00 PM  18:00 UTC
    Oct 8, 2020 02:00 PM  18:00 UTC
    Please download and import the following iCalendar (.ics) files to your
calendar system.

Join from PC, Mac, Linux, iOS or Android:
    Password: 339768

Or iPhone one-tap :
    US: +14086380968,,291630853# or +16468769923,,291630853#
Or Telephone:
    Dial(for higher quality, dial a number based on your current location):
        US: +1 408 638 0968 or +1 646 876 9923 or +1 669 900 6833 or +1
253 215 8782 or +1 301 715 8592 or +1 312 626 6799 or +1 346 248 7799
 or 877 853 5247 (Toll Free)
    Meeting ID: 291 630 853
    Password: 339768
    International numbers available: https://iso.zoom.us/u/abhaIjFKLZ

Or Skype for Business (Lync):


1. Opening and introductions

1.1 Roll call of participants

Andrew Lumsdaine, Phil Ratzloff, luke D'alessandro, Marco Foco, Richard  Dosselmann, Scott McMillan, Michael Wong, Michael Chiu, Kevin Deweese

1.2 Adopt agenda


1.3 Approve minutes from previous meeting, and approve publishing
 previously approved minutes to ISOCPP.org

1.4 Action items from previous meetings

2. Main issues (125 min)

2.1 General logistics

Meeting plan, focus on one paper per meeting but does not preclude other paper updates:

    Apr 9, 2020 02:00 PM: stats paper- DONE
    May 14, 2020 02:00 PM: Stats paper replaces Differential calculus
    Jun 11, 2020 02:00 PM: Graph paper
    Jul 9, 2020 02:00 PM: Stats paper
    Aug 13, 2020 02:00 PM: Differential calculus
    Sep 10, 2020 02:00 PM: Graph paper
    Oct 8, 2020 02:00 PM: stats paper 


all ISO meetings online untilend of Aug

monthly 15th is a deadline

2.2 Paper reviews

2.2.1: ML topics

Richard Dosselman et al

P1708R1: Math proposal for Machine Learning

> std.org/jtc1/sc22/wg21/docs/papers/2020/p1708r2
> above is the stats paper that was reviewed in Prague
> Revie Jolanta Polish feedback.

Freestanding, or accumulator object? Have done both 
series of weight suggested by Michael C
moved to ranges projection as suggested by Jens M
this now looks more compact

latest versoin of python has geometric + harmonic mean following geometric mean
now we can compact all calculation of avg down to 6 , now we are on par with with python has

Yolanta's paper pointed out notion of trimmed mean, implies it is sorted
nice compromise so we don't have to take on the trimmed stats
Luke A: have you considered less strenous restriction of input range as long as you already have forward range?
yes I can change that as I thought foward range was more restricted, I will change that

PR: selection between population and sample, better to use enumeration population vs sample? OK
MF: or do both for when you need to switch to a projection? OK so switch order of projection and population

Quantile as a replacement for median, so you can ask for 0.25 instead of strictly median

have a random_access_range for the sorted case to find the middle position

for mode we can customize with your own allocator, customize what it means to be equal, makes use of output iterator so you can return multiple mode
LA: may want to commit all into ranges, so return mode as a range vs output iterator (view, allowing take_n), so return all of them and not just first
strange that python decided to only return one

PhilR: i was advised too to use output iterator
LA: range's model is composability

PhilR: Also Macro to implementation

accumulator object was more difficult to get right
liek a vehicle that stops every mile, and accumulate it at every mile, user just say they want mean or variance, then pass the mean to this function then you can pull the stats out of these objects; idea is to fuse multipel operations into a single path
so in the end you can do it all in one pass

also have sig variance for a weighted mean

LA: use forwardign references && vs &  aggressively when dealing with ranges

Invite Yolanta, Walter, Eric, and Lisa for July meeting.


Phil Ratsloff et al

P1709R1: Graph Proposal for Machine Learning



Differentiable Programing by Marco Foco

P1416R1: SG19 - Linear Algebra for Data Science and Machine Learning

P1415: Machine Learning Layered list

2.2.2 SG14 Linear Algebra progress:
Different layers of proposal

2.2.3 any other proposal for reviews?

2.3 Other Papers and proposals

2.5 Future F2F meetings:

2.6 future C++ Standard meetings:

-2020-02-10 to 15: Prague, Czech Republic

- 2020-06-01 to 06: Bulgaria
- 2020-11: (New York, tentative)
- 2021-02-22 to 27: Kona, HI, USA

3. Any other business

New reflector


Old Reflector

Code and proposal Staging area

4. Review

4.1 Review and approve resolutions and issues [e.g., changes to SG's
working draft]

4.2 Review action items (5 min)

5. Closing process

5.1 Establish next agenda


5.2 Future meeting

    Jun 11, 2020 02:00 PM
    Jul 9, 2020 02:00 PM
    Aug 13, 2020 02:00 PM
    Sep 10, 2020 02:00 PM
    Oct 8, 2020 02:00 PM