Proposed Agenda for the DECS Brokers Workshop at Lawrence Berkeley National Labs on Thur/Fri Feb 21/22
Important Notes
On Thursday we are in Building 62 Room 203, and on Friday we are in Building 2 Room 100B.
More information about visiting LBNL can be found here: https://www.lbl.gov/visit/
This informal workshop has no funding for catering, so we have planned 30 minute breaks to allow time to go to the LBNL cafeteria for a coffee. Note the LBNL espresso bar closes at 3pm.
Transportation
You should all be receiving an LBL Shuttle Pass. There are two shuttle routes, Blue and Orange. The routes and timetables are found in the following link: https://www2.lbl.gov/Workplace/Facilities/Support/Busses/all_routes.html
Orange: Route goes through the full perimeter of campus including downtown. Route does not go by LBL Guest House. Route goes by Building 62 where we meet on Thursday.
Blue: Route goes through the north of campus and downtown but not the south of campys. Route goes by LBL Guest House. Route goes by Building 62 where we meet on Thursday and Building 2 (the stop marked 54) where we meet Friday.
The Guest House is very close to Building 2 and a 8 minute walk to Building 62.
THURSDAY
Location: Building 62 Room 203
Session 1: Science Vision and Precursor Surveys (9:00 - 10:15)
Goals: To get everyone on the same page (welcome, introduction, work shop goals, “stand up” session), to establish the scientific priorities of a DESC broker, and review lessons from previous surveys.
- Welcome < 10 mins
- Science Standup 10 mins: Really short ~ could be 15 sec a person. Name. Some/all of the following Interest in the topic / Goals in the workshop / involvement or planned contributions / Requirements for science case.
- Talk by Michael Medford “Microlensing with ZTF: Breaking in Spin with Brokers” 10 + 5 mins
- Comments and Discussions on lessons from previous and ongoing transient surveys, and specific science cases led by several people: 40 mins
Break: Coffee/snacks ON OUR OWN (10:15 - 10:45)
Session 2: LSST Prompt Processing Data Products (10:45 - 12:00)
Goals: Overview of the prompt processing data products from LSST and ones currently available from ZTF, and important differences. An overview of the Brokers proposal process. ** (Eric Bellm and Melissa Graham)
-
LSST Prompt data products: contents, expected data sizes, latency of availability, data access mechanisms (e.g., include the LSST alert filtering service). 40
-
The planned LSST Broker selection process (LDM-612), 15
-
Comparison of LSST and ZTF packets, with an emphasis on differences (sandbox infrastructure should be created for LSST alerts, but be testable on ZTF alerts). For example, forced photometry does not exist in ZTF alerts; what else is missing in the ZTF alert packet that DESC would want to be simulating? 15
-
Eric Bellm presents his GitHub for dealing with ZTF alerts (alert packets, avro format, kafka streaming) : 15
Break: Lunch ON OUR OWN (12:00 - 1:00)
Session 3: Broker Components (1:00 - 2:30)
Goals: Understand the building blocks of brokers, essential components for follow-up prioritization.
- Connections between an LSST-DESC broker and machine learning, Emille Ishida (15+5)
- RAPID: Early Classification of Explosive Transients using Deep Learning, Daniel Muthukrishna, 15 + 5
- Deep learning in Astronomy and its application within the ALeRCE system, Guillermo Cabrera-Vives, 15 + 5
Break: Coffee/snacks ON OUR OWN (2:30 - 3:00; note the LBNL cafeteria espresso bar closes at 3pm)
Session 4: Infrastructure in Development (3:00 - 5:00)
Goals: Formal presentations that detail the plans from Broker and TOM developers, including implemented approaches to broker components, choice of API and schema formats. Speakers should please focus on aspects that would be most relevant to a DESC broker.
- ANTARES is Live! Here’s how you use it! 15 + 5
- The Universe in a stream: brokering alerts with the ALeRCE systemThe 15 + 5
- TOM Toolkit and the AEON network: software to make observations and an expanded telescope network to run them on 15 + 5
- MARS: Make Alerts Really Simple, Rachel Street, 5
- LASAIR 15 + 5
Optional/overflow: Preparation for Hands-On Work (TBD, if time allows)
This time might be used for account setup, permissions, a sandbox environment pre-tutorial, software installation, etc. Whatever is needed. We could also spend this time on a hands-on demo, or a general discussion about science objectives and alerts.
Dinner ON OUR OWN
FRIDAY
Location: Building 2 Room 100B
Session 5: Additional Talks and a Group Discussion on “Charting the Course Forward” (9:00 - 10:15)
Goals: (I) There are some people who can only attend on Friday, and some potential topics who still do not have an identified speaker. (II) We will discuss charting the course forward for a DESC broker and its components.
Talks:
- PLAsTiCC, Renee’ Hlozek, 15 + 5
- NERSC, Debbie Bard, 15 + 5
- SkyPortal, Stefan Van der Walt, 15 +5
Discussion on Broker components:
As a group, we will compile a list of needs and make a rough flow-chart diagram for the DESC broker.
Break: Coffee/snacks ON OUR OWN (10:15 - 10:45)
Session 6: Sandbox Setup (10:45 - 12:00)
Goals: Based on the results of the morning’s “big-picture” discussion about a future DESC broker, we will focus on what is needed in terms of a common sandbox infrastructure that can use precursor data and simulations in order to develop the envisioned DESC broker components.
- Spark usage for Broker Components
Break: Lunch ON OUR OWN (12:00 - 1:00)
Session 7: Hacking with Alerts and Sandbox (1:00 - 2:30)
Goals: Face-to-face hack session to initiate sandbox session.
The hacking topics will include, e.g.:
- working together to begin setting up the sandbox environment
- working in small groups to achieve personal goals regarding alerts processing
Break: Coffee/snacks ON OUR OWN (2:30 - 3:00; note the LBNL cafeteria espresso bar closes at 3pm)
Session 8: DESC Broker Discussion and Planning (3:00 - 5:00)
Goals: Reports from the hack session, assignments of action items, and preparation of documentation.
An open discussion session for DESC to answer the following:
- will DESC’s science goals be met by brokers in development?
- should DESC propose a broker?
- what would the DESC broker’s distinguishing features be?
- where would a DESC broker be hosted?
- who will be in charge of the DESC broker proposal during the selection process?
- should we prepare an LSSTC Enabling Science proposal to aid future work?
Preparation of a set of summary slides for presentation during the DESC meeting.