Hi Guys,
I found something that some of you may like: http://procworld.blogspot.nl/
It's a fully procedurally generated world (landscape, foliage, architecture)
Monday, July 2, 2012
Thursday, June 21, 2012
Not everyone received Maarten's email today, so I'll post it here:
---------------------------------------
Hi all,
-------------------------------------
Report Meeting Master AI (19 June 2012)
Present: Maarten van Someren (OC, chair and minutes), Maarten van der Velden (OC), Frank Aldershof, Christof Monz, Gwenn Englebienne, Shimon Whiteson and about 15 students (a mix of first and second year masters, and of UvA and external bachelors)
Entry requirements / prior knowledge:
• Important things are: MATLAB, linear algebra, calculus, statistics
• The schedule leaves little time to catch up with this on the side
• Maybe it is better to start with the maths and then go to the (ML) algorithms, to make sure that students understand this
• Is it possible to have an optional refreshment course?
• Maybe make a little test on the 4 topics, or point to a book / video / exercises so that students can test their level and maybe study a bit before the master starts
• Set up a little test on basics in the first week and offer help to those with problems?
First semester:
• NLP course duplicates Bachelor; this is almost impossible to avoid; maybe accept duplication and waive this for students who can prove that they know the material. Especially in a 3 EC course one cannot repeat and add new material.
• Idea of common / overview is OK
• Level of course in general OK
• Better to set Computer Vision before Int. Multimedia Systems; Int. Multimedia should then be more structured, have more content and depth
• Information Retrieval could be more difficult
• Courses in second semester / track hardly build on first semester
Rest of programme:
• Students would like to have more choice in courses, especially in the second semester; this makes it possible to reach more depth in a topic; these courses can also be elsewhere in UvA, VU or other
• Structure of programme in tracks may not be necessary, perhaps only to explain the main topics in the programme
• Announce that it is possible to do a project in Free Choice credits; is it possible to do a project as track/AI course?
• 8-‐8-‐4 schedule: mostly positive; courses need to be adapted; little room to catch up
• The topics for profile projects should be announced a week before the project starts so that students can choose
Support:
• It would be good if teachers/assistants have a fixed time where students can ask questions (i.e. schedule an office hour a week during the course)
General:
• Some courses provide no feedback on projects/practical work; this type of feedback would be very instructive and useful
• Some exam questions are repeated; this should not happen
• Deadlines for grading are not met
• OC AI should communicate more with students, e.g. by maintaining a webpage on Blackboard with minutes on meetings, room for discussion, etc.
• A meeting like this is quite useful; should be organized on a regular basis
---------------------------------------
Hi all,
Thank you all for your contribution to the
meeting last Tuesday. I think it was very useful, and I got the idea
that this feeling was mutual.
Here is the report Maarten van Someren and me made.
With kind regards,
Maarten
-------------------------------------
Report Meeting Master AI (19 June 2012)
Present: Maarten van Someren (OC, chair and minutes), Maarten van der Velden (OC), Frank Aldershof, Christof Monz, Gwenn Englebienne, Shimon Whiteson and about 15 students (a mix of first and second year masters, and of UvA and external bachelors)
Entry requirements / prior knowledge:
• Important things are: MATLAB, linear algebra, calculus, statistics
• The schedule leaves little time to catch up with this on the side
• Maybe it is better to start with the maths and then go to the (ML) algorithms, to make sure that students understand this
• Is it possible to have an optional refreshment course?
• Maybe make a little test on the 4 topics, or point to a book / video / exercises so that students can test their level and maybe study a bit before the master starts
• Set up a little test on basics in the first week and offer help to those with problems?
First semester:
• NLP course duplicates Bachelor; this is almost impossible to avoid; maybe accept duplication and waive this for students who can prove that they know the material. Especially in a 3 EC course one cannot repeat and add new material.
• Idea of common / overview is OK
• Level of course in general OK
• Better to set Computer Vision before Int. Multimedia Systems; Int. Multimedia should then be more structured, have more content and depth
• Information Retrieval could be more difficult
• Courses in second semester / track hardly build on first semester
Rest of programme:
• Students would like to have more choice in courses, especially in the second semester; this makes it possible to reach more depth in a topic; these courses can also be elsewhere in UvA, VU or other
• Structure of programme in tracks may not be necessary, perhaps only to explain the main topics in the programme
• Announce that it is possible to do a project in Free Choice credits; is it possible to do a project as track/AI course?
• 8-‐8-‐4 schedule: mostly positive; courses need to be adapted; little room to catch up
• The topics for profile projects should be announced a week before the project starts so that students can choose
Support:
• It would be good if teachers/assistants have a fixed time where students can ask questions (i.e. schedule an office hour a week during the course)
General:
• Some courses provide no feedback on projects/practical work; this type of feedback would be very instructive and useful
• Some exam questions are repeated; this should not happen
• Deadlines for grading are not met
• OC AI should communicate more with students, e.g. by maintaining a webpage on Blackboard with minutes on meetings, room for discussion, etc.
• A meeting like this is quite useful; should be organized on a regular basis
Saturday, June 9, 2012
Message about meeting with Theo/teachers
---------------------------------------------------------
Edit: "The Master AI meeting next Tuesday (June 19, 3PM) will be held in room D1.116"
---------------------------------------------------------
Dear Theo, Master AI students,
Edit: "The Master AI meeting next Tuesday (June 19, 3PM) will be held in room D1.116"
---------------------------------------------------------
Dear Theo, Master AI students,
As you've
probably heard, the AI Programme Committee (OCAI) will be organizing a
meeting to discuss the Master Artificial Intelligence, especially the
first semester of the programme.
This meeting will be held on Tuesday, June 19, at
3PM (room to be announced), and will take about an hour, with a
possibility to continue discussion afterwards for the ones interested.
Everyone involved in the master programme is invited.
To get an idea what exactly will be covered, a small agenda follows
that will guide the discussion. It has some rough guides of the time
each discussion will take. Please remember that the goal of the
discussion is to find out what the strong points of the master programme
are, and to get ideas for improvement of the master programme.
Agenda
1. (Missing) prerequisite knowledge for the programme (10min)
2. Programme of the first semester (15min)
a. The concept of a broad first semester and specializing afterwards
b. The relevance of the combination of main courses/subjects covered
3. Qualitative level of the courses (difficult/easy, much or little work), distribution of the study load (15min)
4. Relation of the first semester with the tracks and their courses (10min)
5. Support (10min)
a. Communication with the teachers, programme
b. Communication with / Information from the university and faculty (& Education Service Center)
If
you think an important subject is missing on the agenda, please let me
know, and we will see whether we can discuss that too. If you cannot
attend the meeting, but like to give your input, please mail it to me.
If you know of other master AI students who like to attend, please invite them.
I hope to see you on the 19th,
With kind regards,
Maarten van der Velden
Secretary OCAI
Tuesday, May 1, 2012
meeting with Theo & teachers
During one of the last AAMAS practical sessions we were made aware of some upcoming meeting/discussion between Theo, students and teachers linked to the master AI. We would be able to discuss the master program, whats good/bad, what needs changing etc
I wrote my contact details on the sheet of paper but I haven't heard anything in over a month. Does anyone know when this meeting is going to take place? (maybe I already missed it?)
I wrote my contact details on the sheet of paper but I haven't heard anything in over a month. Does anyone know when this meeting is going to take place? (maybe I already missed it?)
Saturday, April 14, 2012
Game Programming
Does anyone know if the Game Programming assignment (presenting a Scratch game + a retro game) for next week is to be done individually or as group? Its not clear based on the instructions...
I sent the teacher an email on Wednesday but no reply yet.
I sent the teacher an email on Wednesday but no reply yet.
Thursday, March 1, 2012
Tuesday, February 28, 2012
Friday, February 24, 2012
Basis Functions
Hi... If somebody could answer us this:
"What are basis functions, what is the purpose of using them and how can we learn basis functions from data?"
Thank you.
"What are basis functions, what is the purpose of using them and how can we learn basis functions from data?"
Thank you.
Tuesday, February 7, 2012
Geometric Algebra
Hey all,
Last period I taught a bachelor's course on image processing with Leo Dorst. Through his site I found out about this thing called "Geometric Algebra", which is basically a way of doing linear algebra in a much more intuitive and elegant way. Having looked into it just a little bit, it seems to me to be the way linear algebra should be taught. In general I think that a simpler way of doing math will often suggest new methods, that would have looked very complicated in a more convoluted formalism.
I have some ideas about doing object recognition and scene understanding using geometric relations between object parts, which is why I want to learn about GA. I also think this will come in handy in many places like machine learning, given that linear algebra is so pervasive in our field (and any engineering field really). Learning a better way to think about spaces, geometry, etc. seems like it would pay-off sooner or later for any of you.
I've talked to Leo about this, and I'm going to do a 6EC course with him, which basically consists of reading his GA book, doing exercises, some private tutoring and 2 take-home exams. I've heard that many of you are not too excited about the courses this semester, so I thought you might want to join me in this and drop some other course. I've spoken to Tobi about this, and he was enthusiastic about it. Robrecht has already done this GA course and is also very positive about it.
So take a look at the GA website, Leo's website, and if you want to join us let me know ASAP. You will have to buy the book, and Leo will hand out the first chapter as pdf this week.
One last thing to convince you: Leo is a really awesome teacher, and following a small-scale class with him should be really nice.
AAMAS; looking for group
Autonomous Agents and Multi-Agent Systems requires us to form groups for two separate tasks:
- leading a paper discussion
- doing a MAS project
Monday, February 6, 2012
Google's Ngram data
In today's Unsupervised Language Learning lecture, Google's Ngram dataset was mentioned. You can use the Ngram Viewer to see how popular a combinations of words has been throughout history. You can also directly download the Ngram dataset to run your own experiments.
You can also watch the following TED talk, if you find the whole thing interesting:
You can also watch the following TED talk, if you find the whole thing interesting:
Tuesday, January 31, 2012
Resit IMS
The Resit of IMS is next week. Are there results already of the exams ? Or have I not seen it?
Thursday, January 26, 2012
Helpful link for timetables, exam dates etc
Here's a list of all the AI exams with date, time and location for the 2011-2012 year.
https://datanose.nl/#examlist
I'm not sure that will automatically bring up the list for the master AI (it does on my pc but might just be because I've visited the page before), if not then just search for it in the drop down menu.
Also interesting to check out the rest of the datanose website, it contains lots of info in one place. For example, you can see the timetable for a specific room.
https://datanose.nl/#
https://datanose.nl/#examlist
I'm not sure that will automatically bring up the list for the master AI (it does on my pc but might just be because I've visited the page before), if not then just search for it in the drop down menu.
Also interesting to check out the rest of the datanose website, it contains lots of info in one place. For example, you can see the timetable for a specific room.
https://datanose.nl/#
Tuesday, January 17, 2012
Knowledge Representation exam questions
In case, for one reason or another, you don't have access to the exam questions of last November, I scanned and uploaded them. Here goes:
Also, good luck if you will be sitting for this exam!
Wednesday, January 11, 2012
MLPR resit
Yo guys, happy new year :-)
Along with Kristin I went today to Gwenn's office, in order to discuss the problems with the MLPR resit, and especially the point that many of the people intersted in such a resit were not in Amsterdam the day of the resit, because of the last-minute changing of the latter's schedule.
He said it would be possible to arrange a new exam date if the number of the interested students is sufficiently high, so if there is someone in this.. let's keep in touch!
Furthermore, he mentioned that the pass-mark for the written exam is 5, with no roundings (so if you toke 4.9 in the written exam you have to take the exam again).
See you,
Dario
EDIT: We went to Gwenn's office today, reporting that 7 people (Kristin, Cem, Nikolas, Vasilis, Katia, Koen and me) were unable to take the first MLPR resit, asking if it is possible to arrange one more exam date. This turned out to be probably possible: the exam date will be somewhere around the third week of February. He will let us know via email.. Bye!
Along with Kristin I went today to Gwenn's office, in order to discuss the problems with the MLPR resit, and especially the point that many of the people intersted in such a resit were not in Amsterdam the day of the resit, because of the last-minute changing of the latter's schedule.
He said it would be possible to arrange a new exam date if the number of the interested students is sufficiently high, so if there is someone in this.. let's keep in touch!
Furthermore, he mentioned that the pass-mark for the written exam is 5, with no roundings (so if you toke 4.9 in the written exam you have to take the exam again).
See you,
Dario
EDIT: We went to Gwenn's office today, reporting that 7 people (Kristin, Cem, Nikolas, Vasilis, Katia, Koen and me) were unable to take the first MLPR resit, asking if it is possible to arrange one more exam date. This turned out to be probably possible: the exam date will be somewhere around the third week of February. He will let us know via email.. Bye!
Subscribe to:
Posts (Atom)
