Extension in deadline for final report
Hope the exam period is going well.
You may submit the final report until Tuesday 14.7 end of day.
Hope this helps,
|פורסם ב- 12/7/2020, 06:28:35 Created on 12/7/2020, 06:28:35 Создано 12/7/2020, 06:28:35 تم النشر ب- 12/7/2020, 06:28:35|
Online lecture by Sven Konig @ CMU RI
|I want to share with you an online talk by Sven that was given at the Robotics Institute at Carnegie Mellon University.|
Thanks to Tom for pointed it out to me!
|פורסם ב- 20/5/2020, 20:26:02 Created on 20/5/2020, 20:26:02 Создано 20/5/2020, 20:26:02 تم النشر ب- 20/5/2020, 20:26:02|
MAPF @ SoCS
|This is a message I got from Sven Konig.|
Thought worth re-publishing if anyone is interested.
The International Symposium on Combinatorial Search (SoCS) will be held from 26th - 28th May, 2020.
Each year SoCS offers a unique opportunity to hear about the newest advances in the field of Heuristic and Combinatorial Search and to engage with some of the brightest minds working in those areas. SoCS attendees include world-leading academics, outstanding early career researchers and the next generation of leading young scientists.
In this its 13th year the Symposium will take place online and our entire technical program and proceedings are being made available to all, including a larger number of talks on multi-agent path finding.
This year SoCS has aligned itself with the 17th International Conference on the Integration of Constraint Programming, Artificial Intelligence, and Operations Research (CPAIOR). To celebrate we have will have special activities which aim to bridge the gap between the area of Heuristic and Combinatorial Search and that of Discrete Optimisation.
Among the many highlights at SoCS 2020:
- 25 talks across five technical sessions (plus two award winners!)
- Keynote talk from Rina Dechter (SoCS plenary)
- Keynote talk Peter Stuckey (special activity, joint with CPAIOR)
- Master Class sessions (special activity) where leading experts will
disseminate recent developments on a variety of exciting search-based
Every talk will be streamed live on YouTube and will be subsequently made available for viewing on-demand.
HOW TO PARTICIPATE:
1. Follow our technical program from the SoCS homepage: https://urldefense.com/v3/__https://helios.hud.ac.uk/scommv/socs2020/main.php page=program__;!!LIr3w8kk_Xxm!7NqjdfE-WP13KDc7vU8NsxC6z2mDhwG6uIfNAYPOj04JhEG2VtWGzxNunpzQjA$
2. Register for the SoCS Forum, our dedicated community hub for news and discussions at any time: https://urldefense.com/v3/__http://forum.search-conference.org__%3B%21%21LIr3w8kk_Xxm%217NqjdfE-WP13KDc7vU8NsxC6z2mDhwG6uIfNAYPOj04JhEG2VtWGzxPhSgFUOw%24
3. Join our Discord server (details on the forum) for live chat (text/video/voice) with all speakers and other participants.
We hope to see you soon in cyberspace!
Daniel and Mauro
|פורסם ב- 20/5/2020, 09:54:15 Created on 20/5/2020, 09:54:15 Создано 20/5/2020, 09:54:15 تم النشر ب- 20/5/2020, 09:54:15|
Grades for CBS project published
Grades for the CBS projects are published.
In addition, I published a solution under "חומר המקצוע" so you can see the "formal" solution.
A few notes on the grading (given out of 30):
- For submitting the intermediate submissions (task 0-1 and 2) you received 2 points for each task. So everyone got 4 points automatically.
- For successfully answering 2.5 you got 6 points. If you answered all of the subsections correctly and not only one, you got an extra two points.
- For the CBS part (task 3) you got 15 points if the results were correct and optimal.
- For task 4 you got 5 points if the results were optimal.
- Partial points were deduced when the results were not optimal
If you want to discuss your grades and receive more-specific comments on your submission, feel free to contact me.
|פורסם ב- 17/5/2020, 10:01:12 Created on 17/5/2020, 10:01:12 Создано 17/5/2020, 10:01:12 تم النشر ب- 17/5/2020, 10:01:12|
Feedback and presentations available
Student presentations are available under "course material".
Here is some feedback so hopefully other can learn:
Both presenters did a great job. Presentation were visually appealing and clean and they did a great job in responding to the student questions and managing time.
In general, I prefer that the language would be in English.
* Moving Agents in Formation in Congested Environments *
My biggest comment (which is small) regards the notion optimality that caused all of us some confusion should have been clarified.
On a minor technical note - in the explanation of focal search, it was described as an absolute value (8 in your example) but it is actually a relative ratio from the f-value at the head of the queue.
No citation was given for focal search.
Some monior comments:
recognized -> represented / described
4 neighbor grid -> 4-neighbor grid
why  and not ()
Total Formation deviation -> Total formation deviation
* Lifelong Multi-Agent Path Finding in Large-Scale Warehouses. *
I especially enjoyed the description of the empirical results which tried to provide some design considerations (e.g., why PBS outperforms CBS).
Since PBS was shown to outperform other methods, it seems surprising that token-passing was explained and not on PBS.
The presentation containing very little technical information and the presenter relied on his oral presentation to fill in the details.
I think this caused long and vague descriptions resulting in some confusion among the students.
Some minor typos:
token passing algorithm -> token-passing algorithm
requests -> requirements
|פורסם ב- 7/5/2020, 11:01:28 Created on 7/5/2020, 11:01:28 Создано 7/5/2020, 11:01:28 تم النشر ب- 7/5/2020, 11:01:28|
Guidelines for final report
|Recall that the final report's objective is to force you to understand a paper that you did not present.|
This will constitute as 20% of your final grade (after the new point distribution).
The report should be between 1-2 pages and should include:
(1) A summary of the paper in your own words (0.5-1 page).
Try to answer the following question "what is the ONE key insight that enabled the work of the paper?"
(2) Directions for future work? (0.5 page)
What is missing, how could the work be extended?
Design in detail a project around the paper (1 page)
What are the outcomes (implement the approach is not a good project)?
What are the expected results?
What tools are required to implement the project?
|פורסם ב- 7/5/2020, 09:28:13 Created on 7/5/2020, 09:28:13 Создано 7/5/2020, 09:28:13 تم النشر ب- 7/5/2020, 09:28:13|
Talks at AAMAS 20
The International Conference on Autonomous Agents and Multi-Agent Systems will be held online (via recorder talks) this year free to all.
The talks we heard yesterday will be presented there.
The first one is already available at
If you are interested in the other one, it will appear in the upcoming days at
In addition, there is an interesting paper,
Multi-Agent Path Finding in Configurable Environments by
Matteo Bellusci, Politecnico di Milano
The talk is available at
I encourage you to check out the talks and any additional ones you find interesting.
|פורסם ב- 7/5/2020, 09:19:26 Created on 7/5/2020, 09:19:26 Создано 7/5/2020, 09:19:26 تم النشر ب- 7/5/2020, 09:19:26|
Changing grade distribution
|I understand that Task (4) in the CBS project was much more time consuming than expected.|
To accommodate for this, the final report will be smaller and will constitute only 20% of the grade.
The CBS project will be 30% of the grade with the fourth task being only 5 out of the 30 points.
This will allow students who did not complete Task (4) at all to still get a very high grade (95) and to reduce the workload at the end of the semester.
Looking forward to hearing your talks,
|פורסם ב- 5/5/2020, 08:50:43 Created on 5/5/2020, 08:50:43 Создано 5/5/2020, 08:50:43 تم النشر ب- 5/5/2020, 08:50:43|
CBS project submission date
|The submission date for the CBS project was postponed by a week to May 10th.|
|פורסם ב- 30/4/2020, 07:58:40 Created on 30/4/2020, 07:58:40 Создано 30/4/2020, 07:58:40 تم النشر ب- 30/4/2020, 07:58:40|
Testing for task (3)
In task (3) it states that you need to test your code on file `exp3_1.txt'.
This file does not exist.
Please test it on `exp2_1.txt'
|פורסם ב- 14/4/2020, 11:26:54 Created on 14/4/2020, 11:26:54 Создано 14/4/2020, 11:26:54 تم النشر ب- 14/4/2020, 11:26:54|
Reading material for upcoming class
Due to some requests I updated the required reading material for this week's meeting.
You are only asked to read the CBS paper (reference  in our spreadsheet) and you are asked to read the paper on disjoint splitting (reference ) for the following class meeting.
|עדכון אחרון ב- 28/3/2020, 12:56:38 Last updated on 28/3/2020, 12:56:38 Последняя модификация 28/3/2020, 12:56:38 تمت الحتلنة الأخيرة ب- 28/3/2020, 12:56:38|
|I opened a "HW" for the submission of task (0) and (1).|
For the intermediate submissions, please submit just the code.
This is just a sanity check that all students are all on the same page.
I will give 0 or 2 points for each intermediate submissions (on dates 29.3, 7.4 and 26.4) .
The real grading (rest of the 14 points out of 20) will be given when the mini-project / HW is submitted.
Here you will have to attach a small text describing your implementation, answering questions (e.g., 1.5 that we discussed yesterday) with some screenshots.
|פורסם ב- 26/3/2020, 07:08:33 Created on 26/3/2020, 07:08:33 Создано 26/3/2020, 07:08:33 تم النشر ب- 26/3/2020, 07:08:33|
Lesson 1 - summary
|Hope you all enjoyed our first lesson yesterday.|
For those of you that did not find the spreadsheet, here is the link again:
The tasks for this week are:
(1) Read papers [1-3]
(2) Review A* if needed (e.g., using [4-6])
(3) Implement Task (0) and (1) in 
Where the refs are:
 Hundreds of Cooperative, Autonomous Vehicles in Warehouses
 AI buzzwords explained: multi-agent path finding (MAPF).
 Overview of Multi-Agent Path Finding (MAPF)
For next week, please post questions / problems on Task (0) and (1) by 24.3 using the same spreadsheet (there is a dedicated tab).
|פורסם ב- 19/3/2020, 08:58:40 Created on 19/3/2020, 08:58:40 Создано 19/3/2020, 08:58:40 تم النشر ب- 19/3/2020, 08:58:40|
Corona virus and seminar - update (2)
|Following the fact that we are forced to meet virtually and not in person,I decided to slightly change the seminar's format. I will discuss this in detail on the Wednesday's zoom lecture but the high-level summary is that I don't think it will be effective to have students give a two hour-lecture. Instead, we will spend the first part of the course doing a mini project that will give us hands-on experience with MAPF algorithms.|
Student lectures are postponed to after Independence day and will span only one academic hour per students.
I know these are difficult times and uncertainty is high. Feel free to contact me personally regarding any issue and I will try to respond as soon as possible.
|עדכון אחרון ב- 17/3/2020, 06:47:02 Last updated on 17/3/2020, 06:47:02 Последняя модификация 17/3/2020, 06:47:02 تمت الحتلنة الأخيرة ب- 17/3/2020, 06:47:02|
Corona virus and seminar - update (1)
|Amidst all the confusion, we will attempt to start the seminar next week, as planned, in an online platform.|
This will be done via Zoom (https://cis.technion.ac.il/central-services/zoom-video-conference/)
Please make sure you know how to use Zoom, and login into your account (either your Technion account, or a private one).
The link to the zoom room is -
Meeting ID: 609 146 238
You are welcome to attend the class even if you're not yet registered to the course.
Feel free to email me if you have any questions.
|פורסם ב- 13/3/2020, 12:32:29 Created on 13/3/2020, 12:32:29 Создано 13/3/2020, 12:32:29 تم النشر ب- 13/3/2020, 12:32:29|
Corona virus and seminar
|As you know we are all uncertain regarding the next few weeks due to the outbreak of the Corona virus.|
Currently, the restrictions we have do not preclude us from having lectures with less than 100 students so we will meet as scheduled next Wednesday (13.8) at Taub 8.
However, we are planning for the case where we will have to give the course virtually. If this will be the case, we will use Zoom to give the lectures, and discuss them. Please make sure that you have access to a computer with a good microphone and a webcam.
I will keep you posted once things clear up.
|פורסם ב- 12/3/2020, 11:16:23 Created on 12/3/2020, 11:16:23 Создано 12/3/2020, 11:16:23 تم النشر ب- 12/3/2020, 11:16:23|