Hacker News Reader

Comment by ttub | Hacker News Reader

No offense intended, but what is your background?

I'm a researcher with a physics/stats PhD, and if a colleague approached me and said "stochastic gradient algorithms" entails three highly complex areas of scientific knowledge, I would have been stunned and assumed an undergrad with an English major had stumbled into our lab.

Just because you find something extremely challenging, doesn't mean it is inherently challenging. Considering what a lot of people in my field is struggling with, your example is absolutely trivial. You might want to adjust your ego downwards a bit.

View on HN

No offense taken. I feel humbled. Indeed, your argument supports my view.

I have deep and very high regard for the people who are able to apply ML to fields like DNA analysis or NLP which can take the dreaded "Turing test".

I stand nowhere in the ML arena, but I had tried once and got a good shock of my life: how hellishly difficult the ML can get and how quickly. I really feel humbled. If anything, I learnt to appreciate the width and depth of human brain capabilities. It seems entirely magical now to me how on earth does my brain process/understand such complex things like this very paragraph. Prior to some exposure to ML, I couldn't have appreciated this thing.

>>Just because you find something extremely challenging, doesn't mean it is inherently challenging.

Agreed. I never claimed it anyway. But the kind of problems, for which ML is being applied, the state of the art existing "analyzable algorithms" (like, finding approximate near-optimal solutions for TSP) are far from trivial. In addition to this, we must realize that the ML solution must "beat" these algorithms hands-down in "non-trivial" cases. All this makes ML extremely difficult.

I agree that for real world (and not necessarily state-of-the-art) ML applications, you have to handle many more fields in addition to these 3 fields. All I say is even these three things, when taken together, are very complex things to handle.

edit: typo

I know a simple upvote may (and perhaps ought to) suffice instead of writing this, but I feel compelled to comment on how impressed I am that you responded to "You might want to adjust your ego downwards a bit" with such class and humility. Truly refreshing, truly commendable. I'm using it as a learning experience, because my initial reaction on reading that comment was highly negative and I sincerely doubt whether I would have been able to summon the kind of response you did.

On a morning where I woke up feeling anxious and worried after the events of Monday, this was a small but appreciated little reminder that there's still hope.

I deeply respect the humility of your response, even though I personally would have responded with a simple downvote.

I was curious about what you do, so I visited your HN profile to find more details. As a HN user - and a Muslim - I believe that it would be more useful for yourself and others if your profile listed something about you and where to contact you, rather than a wall of text arguing for why Islam is evil.

Anyways, I hope you have a nice weekend :)

Strictly as a thought exercise, consider how often you find yourself saying "No offense intended."

The person you responded seems like most of "us" - most of us aren't ML experts. He humbly shared his experience with dipping his toe in the ML pool and reported back that he (metaphorically speaking) had to chip the ice off the pool before he could even get in.

This was useful information to "us" (non-experts interested in learning about ML). Calling his difficulties "trivial" for an expert in the field while true, came off as condescending. Telling him (him!) to check his ego came off as egotistical.

No offense intended. ;-)

If you hadn't included that last sentence I wouldn't have downvoted your comment. You can't just say "no offense intended" and then insult someone. Anyway, most of the people reading this are probably regular software developers that are interested in ML, not PHD holders in any field. It's obviously not written for academics at the leading edge of the field. And if something takes several years of dedicated study to master, it's challenging.

I didn't find the parent's comment egotistical at all. I believe they were just commenting on the three different fields that combine to form stochastic gradient descent. The fact that it's viewed as fundamental or trivial only supports his point that things in machine learning are hard, and may require wide knowledge breadth.

I understand how it would look this way to you, but you have to realise that these areas only seem simple to you because they're exactly what you've spent a large part of your life studying.

It's like you're a senior Navy pilot, and you hear a crop duster pilot saying that the Osprey is difficult to fly because you need helicopter piloting skills, plus multi-engine fixed wing piloting skills, plus experience landing on a carrier. He's not wrong, it just doesn't sound that bad to you because you just happen to specialize in the exact combination of skills required.

He didn't say he found it personally challenging. He said this part of optimization finds itself at the intersection of three deep fields of human knowledge. That is undeniably correct.

Thats like saying bubblesort lies at the intersection of combinatorics, topology and complexity theory. True, but pretentious

I don't think it was pretentious, and your analogy is quite off. You don't need to understand any of those fields to grok bubblesort (algorithmics is enough). You do need to know some statistics, calculus and algorithmics to understand stochastic gradient descent.

3 fields, but each only at an introductory level.

you don't need to understand statitstics. calculus (calculus is a high school subject and is hardly deep), and what is "algorithmics"? If you mean the study of any and all algorithms, then sure,

Right, but in this particular case the "stochastic" and "algorithm" parts are trivial. For example, you don't really need to draw on CS/algorithm knowledge to implement or understand gradient descent.

By the time you have a PhD a lot of problems may seem simple, especially in your field, but you spent 6 years of study most people haven't. The truth is that machine learning is the intersection of statistics, data analysis and software engineering. This makes it hard to learn, and also hard to get a job as companies are looking for people who are experts in all 3 fields.

I don't know what OP means by saying SGD. If he/she means it is Gradient-Based optimization algorithm as a whole, it is definitely a challenging and open question, and a lot of researchers trying to improve them.

I understand stochastic gradient descent just fine, but you need to reverse your thought here. Just because something is trivial for us doesn't mean it lacks inherent difficulty. After all, if it was inherently trivial, pretty much everyone would be able to actually derive and implement the algorithm, which the vast majority of the human race can't.

Congratulations - you know something about your field. No need to be a dick about it.

I am sure there are fields where you don't know stuff, so please grow up.

Reply on news.ycombinator.com