The Singularity

nbro

01:03

The following question should be closed as off-topic.

2

I'm working on a q-learning project that involves a "robot" solving a maze, and there is a problem with how I update the Q values (every time the robot ends up switching between two squares instead of actually learning) but I'm not sure where: I am at my wits end. Any pointers are welcome, here i...

AI Meta Questions

05:16

0

Q: What should the on-topic page look like?

Similarly to What should the AI.SE Site Description be? and after the discussions On-topic and off-topic pages need to be clarified and Who decides and writes the on-topic and off-topic pages?, I think it is time to vote for a clearer and updated version of the on-topic page, which users (but esp...

discussion scope site-promotion

DuttaA

09:37

@nbro I appreciate your enthusiasm but I can't agree with it simply because beggars can't be choosers. Our stack is just going by due to the benvolence of few users

Coincidentally when one expert is busy generally another expert becomes available

So if we close low tier questions which can somewhat be answered by intermediate users we risk losing interested contributors with potential to contribute in the future.

Manuel Rodriguez

11:43

@nbro What is the desired close percentage? Each user has the ability to vote for close or leave open a question. Instead of arguing with the question itself, it make sense to define what the target corridor is. Is a close percentage of 0% acceptable, which means that no question at all are put on hold and defined as offtopic, do we like a ratio of 10% which means that sometimes a question is closed or is the wanted ratio 30% which means, that very often an OP is rejected.

nbro

14:07

@ManuelRodriguez I don't think that we need to define a prior percentage of questions we need to close. We need to define clearly the scope of the website and strictly adhere to it, otherwise, this website is just a bunch of random questions.

@DuttaA By "low-quality questions" I don't mean basic questions like "What's an agent?", which are well-accepted on this stack, even with my propose new description of the on-topic page, but I mean questions that are poorly written, unclear or off-topic. We cannot lose 1 hour in trying to understand a question, which the OP did not put any effort into writing. In all SE sites, people expect the asker to put effort into writing a clear question and to do a little bit of research before posting.

We are volunteers. We should dedicate our effort to the questions asked by people that also put some effort into writing their questions. More importantly, we should close all questions that are off-topic, unclear, too broad and primarily opinion-based. I don't care if we have only 5 questions per day. I prefer having a few high-quality questions than having thousands of low-quality questions.

I even extended my description of the on-topic page because someone of you believes that we should deal with programming questions.

If most users are not willing to be more strict and adhere to clearer guidelines, I will not run for a moderator position and I will stop visiting this site. I am tired of cleaning up questions because people do not put any effort into writing them. I am tired of seeing off-topic questions not being closed. And I am tired of trolls.

DuttaA

@nbro well on the contrary you should run for it if you have a vision for the stack. I am not a prolific contributor so my opinion is not that important.

nbro

14:26

I think that the opinion of the regular users is quite important, otherwise, I would not be asking questions on meta, so your opinion or perspective is clearly important because you are one of the few regular users.

If we want to go to the next level (i.e. graduate), we need to have a clearer idea of the goal of the stack and we need to do our job diligently (which includes closing questions when they need to be closed)

It is time to stop being nice when people do not really want us to be nice, because they show no effort. It takes effort to write an answer, so we expect a similar effort to write a question.

With this, I don't mean we should be rude. Of course, we should try to ask for clarification and explain the purpose of the site, if the user is apparently unaware of it. However, if the asker is not willing to clarify his question, to understand that multiple questions in the same posts are not very appealing, etc., then there is no point in losing more time with this asker.

Manuel Rodriguez

@nbro If i have understood your concern right, you see many questions which should be closed by the user voting system. The reason why these questions aren't closed, is because other users have a different definition of the term offtopic. User1 has pressed the close button, but user2 has pressed the “leave open” button. The question isn't, if a question fits to SE.AI but how to convince other user to judge in that way. In the Wikipedia universe the problem is discussed

under the topic exclusionist vs inclusionist.

nbro

@ManuelRodriguez Well, this is another reason why I think regular users should decide on a new, clearer and more specific on-topic page. Also, note that, sometimes, certain questions have multiple problems, so you should not be surprised when they are flagged for multiple reasons.

By the way, I believe that my proposed on-topic page can still be improved. We need to clarify further which programming, software and hardware questions are on-topic

I am not happy with "Hardware or software used in the context of AI (e.g. "Which hardware did Google use to train model X?")"

We should also clarify which academic aspects of AI are on-topic. I think we should just demand the questions to not be primarily opinion-based.

Manuel Rodriguez

14:47

@nbro It sounds, that the problem which question should be closed and which not will remain a hot topic in the future. As far as i can see from the Meta forum, there is a dedicated tag (scope) which is reserved only for such conflicting opinions. I'm relative sure, that the users in SE.AI are not able to agree on a shared position, so it will become exciting in every single close-voting.

nbro

@ManuelRodriguez Well, maybe we should wait to update the on-topic page until we have an answer which has been upvoted 5 times (which is the number of votes needed to close a question, unless a voter is a moderator).

I hope will not need 3 years to update the on-topic page!

DuttaA

16:37

@nbro at earlier times (when you were not active here) I also had raised some flags and close votes, but most of them are struck down (sometimes reasonable sometimes for unreasonable) so unless the question is really bad I generally don't close or flag it

But yes with Ana active moderators I might close or flag more (since delays make me forget of the outcome, I.e. whether my flag was useful or not)

Also o can't quantify what's on topic and off topic at a single go. All I can do is determine whether questions are on/off topic (as per what I think suitable). Since some user will always come up with something new which is on the brink of on/off topic

nbro

@DuttaA I completely agree with you that, at least during this year, moderators have not done their job well and have not been active. Furthermore, they have almost never enforced the rules of the site, not closing almost any question. That's why I wanted a change in moderators. I think moderators should be the most active users who have a clear vision of the purpose/goal of the site.

@DuttaA Again, this is another reason why I think we should have clearer guidelines. The current guidelines are ambiguous, so, of course, you and I (and many others) do not know what this site is for or when to flag a question for closure.

In my proposal, https://ai.meta.stackexchange.com/a/1616/2444, I think it will be particularly useful if users agree with the Notes sub-section that I've written. I've written these notes so that to increase the probability of receiving more high-quality questions, which does not mean that they will all be high-level (they can also be basic questions). It just means that we should receive fewer questions of the form "How does A* work?".

Or questions like "Do you like more Python or Java for AI?", which are in general not suited for SE sites, that is, they should be closed as primarily opinion-based. As I say in one of the notes.

> Ideally, we are looking for questions that can be answered objectively. More precisely, do not ask for advice (such as career path recommendation or a tool, which are, in general, off-topic here anyway) but for facts and arguments

Manuel Rodriguez

17:47

@nbro quote "Do you like more Python or Java for AI?". Java is better, because a stack-based language can be created easily with automatic programming while a procedural language code is much harder to synthesize from a description in an ontological format. Orlov, Michael, and Moshe Sipper. "FINCH: A system for evolving Java (bytecode)." Genetic Programming Theory and Practice VIII. Springer, New York, NY, 2011. 1-16.

nbro

@ManuelRodriguez Can you provide a link to the pdf and specifically state the section you're quoting? Furthermore, can you explain how your comment is reasonably related to my comment? Please, stop posting out-of-context content! or at least stop posting unclear quotes which are apparently unrelated. This community would benefit a lot if you stopped posting apparently spam questions and answers.

You should vote to close the following question as TOO BROAD.

-1

Q: How do I change the values of a neural net

I'm trying to have a go at building a neural net, but I can't seem to figure out how to optimise the connections. I've tried to have a look online and it came up with "backpropagation". I looked through some pages about it, but I can't seem to understand it. It seems to be where you decide on a...

neural-networks ai-basics python backpropagation

nbro

18:10

The following question should be closed as PRIMARILY OPINION-BASED.

-1

Q: Ideas for a thesis on policy gradient methods

I've looked into policy gradient RL the last few months. As I find the topic quite interesting, I've been readings lots of papers about it. My aim is to write my master thesis in Maths about it. I already started out, the preliminary title being "Techniques for variance reduction in policy gradie...

reinforcement-learning academia

Similarly, the following question should be closed as PRIMARILY OPINION-BASED or as UNCLEAR.

0

Q: Ideas for a computer vision project

I'm working towards getting into a Computer Vision-based tech team in college. To get in, I have to impress my college seniors with an independent project. The team works on autonomous driving, so any ideas for a project related to this field would help a lot. I've completed the Andrew Ng course...

computer-vision academia

nbro

23:59

You should close the following question as a duplicate.

3

Q: Why do we need learning in unsupervised learning?

I am not clear with the concept that an unsupervised model learns. We are giving an input and output to the supervised model, so that it can generate a particular value, pattern or something out of it which can be used to categorize something in the future. By contrast, in unsupervised learning, ...

definitions unsupervised-learning

Transcript for