anjc

Hm, not sure. We can easily resolve the structural detail within each grid, so does our brain just interpolate for colour? I wonder would a green grid on the sky make us perceive a green sky, or do we interpolate it here because we know the sky should be blue.

Well, it is B&W. The N here is so large that you can easily resolve the pixels that should be B&W, yet they aren't.

Well why don't we see a B&W image with coloured lines on it? Rather we see a coloured image with coloured lines on it. It's pretty spooky to me.

It's interesting zooming in, until the point that your brains perceives the grid properly, and a B&W image pops forth.

You and everybody replying to you are making the same baffling mistake, i.e., assuming that technical/song writing ability/education = star. If you want technical ability, there are 8 year olds on YouTube with better chops than EVH. Possibly every 8 year old has better chops than Kurt Cobain. But technical ability does NOT make you a rock star. Same applies to the other fields you mention, painting etc.

I'm sure young rock stars (in the same vein as EVH, Axl, Hendrix, Cobain, Morrison) would still exist as the personalities and talent are still there, but rock is not mainstream anymore, so they can't exist. The closest thing to real rock stars now are the likes of Justin Bieber.

You should re-listen to the Beato/Satriani clip you told someone else to listen to, because he essentially says this. The artists you suggest are irrelevant and can not be compared to EVH in any way whatsoever. Guthrie Govan?! Van Halen drove mainstream culture. Nobody outside of guitar forums know Guthrie Govan. The other suggestions in this thread of comparators to EVH and Chris Cornell (Tosin Abasi, Petrucci, Belamy) are just embarrassing.

F1 isn't Indycar. The headline doesn't mentioned Indycar.

This is what demo papers are for. Submit to a demo track, and don't take negative reviews personally :)

> Your intended audience will understand a point, and want you to go further. But the referees are often not the intended audience.

How do you know this? Like both you and OP, I've had my writing misinterpreted by reviewers. These reviewers may not be experts in my precise niche, but reading research is part of their profession. If even they misinterpret it then it is likely I was not clear enough. I have no reason to believe that experts in my niche would not also misinterpret it, and there's no way to verify this once the work is published.

Furthermore, vocabularies and terminology in a field can change. I find it very confusing when a paper from today and 15 years ago are using differing definitions for the same term.

As such I think defensive writing is a good practice, and I'm not sure why making your work more readable is ever a bad thing.

> when you would really just like to go into details on what's cool about the new type of mask, for people who actually care about better masks.

Defensive writing wouldn't stop you from doing this. It would ensure that a wider audience (across fields and into the future) can benefit from your proposals

Although I sympathize with the author, I believe most academics are aware of/experienced all these issues after their first year of PhD. Becoming a professor these days is hyper competitive, you don't accidentally become a prof, or become one by default. So it's a pity that this person would quit over issues that they were aware of from early on.

Regarding papers that shouldn't be papers because it's a waste of resources, I think it's unkind/elitist for someone who rose to a prof position via these resources to then propagate such a sentiment. Only experts and geniuses should attempt to publish? Only people from certain institutions? Why should an author self-censor their own work because one person who chose to be a PC member, to pad their CV (which OP also criticises), doesn't think it's worthy?

In general if people are making poor contributions, their careers in academia will be short-lived, and their individual draining of resources will be minimal. If they get published then it's likely on merit.

Perhaps it's a regional thing then, in my country 2x4's aren't exactly cheap if you're buying enough to build an e.g. woodworking bench, and you need a planer even if they're pre-planed.

I'm currently planning a studio desk right now, funny enough. I could buy one for maybe €500, but just the heavy duty shelf sliders I need will be €70 for a pair. A pre-bought one will also have a light aluminium frame, whereas mine is going to weigh a ton with steel tubing.

Like you say though, if I make one it will fit the space perfectly etc. All in all it's worth it for me even if I personally don't save money

You're comparing apples and oranges, if you're building a desk out of some 2x4s and a spare chunk of wood, and comparing that to $1000 products. Undoubtedly the manufactured products are lighter, more robust, and made from better materials. That isn't to say that these desks are more suitable.

One could build a desk from even better materials than the $1000 desk; maple, walnut, mahogany, stainless steel, brass, whatever. But it wont be <$1000 for materials then, let alone <$100.

I'm in favour of DIY and think it's admirable that you would build your son's desk yourself rather than buy, but I think it's important to be realistic about the costs and materials. These costs were a surprise to me at least, after getting into woodwork/metalwork.

> It's less expensive than pre-made furniture.

Not a chance. Perhaps if you live in the Pacific Northwest, but outside of certain regions wood can be very very expensive.

> I could definately replicate what this guy did if I had the time and money

Perhaps you can but for anyone else reading, this is a big and complicated project. Far beyond the capabilities of an enthusiastic/studious beginner. Your mitre saw or cross cut jig/sled being just a few thousanths off will mess this project up. This guy did a really great job.

Wood's expensive. High quality exotic wood can be very expensive. Someone making their first pieces of furniture will likely (should) be using MDF or pine, same as an Ikea item, but will still be paying more just for the materials than the finished item.

Well you're an extreme outlier. Typically people want age-appropriate content in a language that they speak, or, which at least has subtitles. As you point out, if this is your preference then you can use search functionalities for this purpose.

Surely you can see that, as you are unable to speak every language in the world, yet you understand all media presented to you, that ALL of your media is already implicitly sorted by geography and language?

If e.g. YouTube recommended you the best items from all countries, you'd have a very bad time and would be unable to understand almost all media.

The same applies to age and gender. If you were only recommended items relevant to 90 year olds or 9 year olds, you'd have a very bad time on YouTube etc. Just because you don't notice that your media is implicitly sorted doesn't mean it isn't happening.

> I'm not sure why anyone would want a recommendation system that looks for age/gender/race as that seems fairly irrelevant to interesting content

You can't understand how age/gender/race is important for finding relevant content?! These are extremely important, fundamental, attributes that underpin good recommendations. I.e. they're so important that they generally go without saying, and are baked into the corpuses (e.g. your local news site doesn't recommend articles from another country, Disney's streaming service doesn't recommend knitting shows to children, etc)

Would you be happy if your Amazon Prime or Netflix account only showed items that are interesting to people aged {opposite age group to you} in {country on the other side of the world to you} who are {not your gender}?

I think you'd very quickly be surprised at how much media exists in the world and would quickly realise that you're mistaken.

> ... where algorithms are culture dependent and where not. Low level algorithms like quicksort definitely are not ...

Aside on this. In a 1st year CS class I was surprised to learn that half the class had one method to count the number of days in a month (30 days hath September) and the other half had another (using their knuckles). The method a person knew depended on where they grew up, and neither group had heard of the other method. So at least some search algorithms are culturally dependent :)

> You can do the same thing with comments that he claims TT does with video (dwell time threshold feature)

>I don't work at Twitter, but I doubt that matters very much.

>Also regarding "negative sentiment" feedback from users. TikTok does not have a UI tool to say "I don't like this", but other apps do. For example, you can ask Instagram to "show me fewer posts like this".

There's a big difference in the usefulness of implicit and explicit feedback for recommenders, and there's a long history of research trying to address the shortcomings of each, or combine them to avoid their weaknesses (e.g. research into presentation biases, augmented matrix factorization, MF with implicit data, ensembles)

Implicit feedback (dwell times etc) are particularly prone to noise and biases. On the other hand, the amount of explicit data (likes, downvotes) companies can get is so small compared to implicit data, that many companies don't even bother with their explicit data. E.g. even YouTube claim not to use likes/dislikes, in order to better recommend to the long tail of items/users. So it very much does matter, being able to get good implicit data.

The interesting thing implied by OP is that TikTok seems to have completely sidestepped the pitfalls of each form of data. They in essence seem to have made implicit data be a gold standard judgment of relevance, which is really cool I think.

> I don't want to sound overly dismissive, but it's pretty clear reading this that the author has not worked on recommendation or personalization systems recently.

This is something that hasn't been done in any other large company's recommender system as far as I'm aware. It seems like a paradigm shift to me, albeit, is only possible because of the addictive nature of the app. So at least for similar UIs, this concept makes a lot of recsys research irrelevant and is highly novel. As such, I don't think you're giving OP or TikTok enough credit.

RecSys 2016

Covington, Paul, Jay Adams, and Emre Sargin. "Deep neural networks for youtube recommendations." Proceedings of the 10th ACM conference on recommender systems. 2016.

Nvidia furthermore implemented/released it for TensorFlow, and for their new Recsys engine:

https://ngc.nvidia.com/catalog/resources/nvidia:wideanddeep_...

I'm sorry I don't know how to turn it on. When I go to youtube.com, at the top it has about 15 categories that it has gleaned from my use (e.g. Guitar, Flight Simulators), some of which I'm surprised it learned, as well as All, Mix, Trending. I use Edge and I'm logged in to YouTube.

> Then, if a user watches a single video on a topic outside of their usual interest scope, YouTube replaces any semi-relevant recommendations with with that new category, regardless of if the user is actually interested in the topic.

That's the way it used to work, now it adds that new category to the list of categories it thinks you like (top of the home page), and you can filter or include that category.

This criticism shouldn't apply anymore. They predict categories you like and you can select a category. The 'All' page should be a mix of all categories, not just cat videos. Amazon on the other hand...

> Because YouTube doesn't care what you want--YouTube cares what it can monetize.

> You are the product, not the customer.

> Once you internalize that, stuff that Google does makes sense.

They published their algorithm. They rank items according to predicted watch-time. The other poster is probably correct, in that many people use these long classical music videos as background music over and over (i.e. high watch time) and so they're highly ranked.

If they used to recommend a good mix of content, what do you feel is the issue with recommendations now? Not enough variety? Not enough relevance?

Since when, do you think? I personally think that they're better than ever. On the other hand, I think they moved to more deep learning approaches over the last 3-5 years (based on their research output at least) so perhaps recommendations are perceivably different since then

If they were only recommending comedy from the same artist you may then have the same issue as another comment, which is that recommendations aren't novel enough.

I'd take your point if they only recommended bass videos, but presumably it's just one recommendation in a list of many, with the point being to explore your preferences while also satisfying your demands

I'm always mind-blown reading comment threads like this regarding recommendation, with people annoyed that explicit feedback helps recommenders, or exasperated that they're recommended one item that they haven't approved while also wanting novel recommendations.

1000s of hours of videos are uploaded every minute. 5+ billion videos are watched daily. Are you not rather quite pleased that YouTube can deliver highly relevant top-20 recommendations within milliseconds from such a large and changing corpus, while keeping track of your past preferences, your interactions, your social circle's preferences, popularity and trends, content features, etc?

It's pretty amazing to me. I think people expect recommender systems to read minds.

> "use YouTube infrequently"

> "binging random content"

gets recommended random popular content

> "why are recommendations so bad"

They can't read your mind, you know

Yep. They've had version control for at least a decade, diff'ing also, by way of Compare. I'm also not sure why people are fascinated with using git here. It's weird seeing all of the complex solutions in this thread for a problem which does not exist.

Edit: I meant 'fascinated with using git here in this context'.

> Its led to years of teenagers with no training putting things on their cars

Untrained people fiddling with their cars didn't start with The Fast and the Furious! If anything, this era marked the beginning of the end of all that.