anti-sub capability advancement of PLAN.

The Soviets also developed non-acoustic tracking technologies along with sonar tech, which are areas of research that are relatively underdeveloped. It's much like how sharks use a multitude of sensors besides acoustic senses compared to dolphins.

RedMercury said:

Detection using sonar is a prototypical machine learning problem; judging from the proportion of Chinese grad students working on such topics, the above claims of Chinese inferiority in sonar systems are wrong or will be wrong in the not too distant future.

Click to expand...

Hmmm, can you supply us the source of your opinion please:coffee:

RedMercury said:

Detection using sonar is a prototypical machine learning problem; judging from the proportion of Chinese grad students working on such topics, the above claims of Chinese inferiority in sonar systems are wrong or will be wrong in the not too distant future.

Click to expand...

dude, making high tech military equipments is not doing class project in graduate school. And machine learning itself is still a dummy thing that only applicable for very limited domain. And and, sonar is not a prototypical machine learning problem.

Talking about machine learning, you know sparse data problem? NP complete problem? These always occur in detecting using sonar.

szbd said:

Sonar is much harder than radar. The main problems are

1. sound is much slower, 340m/s, while electromagnetic wave is 300000000m/s
2. sound in water almost never travel in strait line, while electromagnetic wave in the air always travel in strait line.
3. The disturb of sound in water is much stronger than electromagnetic wave in the air.

Therefore, with a bad radar, you can often get something, but with even a good sonar, you often get nothing.

Click to expand...

You forgot to mention that sonar travel speed changes with water temperature, density, alkalinity and salinity. Its not a constant like the speed of light, which is with radar. The speed of travel from source to target and back again is what is used with both to determine range. And if speed varies, then so is the accuracy of range determination.

szbd said:

dude, making high tech military equipments is not doing class project in graduate school. And machine learning itself is still a dummy thing that only applicable for very limited domain. And and, sonar is not a prototypical machine learning problem.

Talking about machine learning, you know sparse data problem? NP complete problem? These always occur in detecting using sonar.

Click to expand...

Dude, there is NO "class projects" in graduate school. ALL graduate students work on real-world problems. You see, unlike undergraduates who pay to go to school, graduate stduents get paid. In order to support graduate students, professors have to obtain funding. So the projects that the students work on have to be significant enough to convince people to give them money. Would you donate a million dollars to a "class project", or would you more likely to donate to someone who is curing cancer?

In fact, I would go as far to say that graduate students are, most times, even more knowledgable than their mentors on many aspects. Based on my own experience as well as others', most of the new scientific discoveries are made by graduate students. Although the credit goes to the professors, they usually stay in their office and write grants and go to meeting to brag about what their students have found. Many of them don't even have a clue as to what their graduate students are doing. It is the graduate students who usually plan and execute experiments and make new discoveries.

vesicles said:

You see, unlike undergraduates who pay to go to school, graduate stduents get paid. In order to support graduate students, professors have to obtain funding.

Click to expand...

Over here, it is the projects that get paid rather than students.

vesicles said:

In fact, I would go as far to say that graduate students are, most times, even more knowledgable than their mentors on many aspects.

Click to expand...

Agree. It is sometime fustrating when a professor pushes you to do something that you know is not going to work.

Engineer said:

Agree. It is sometime fustrating when a professor pushes you to do something that you know is not going to work.

Click to expand...

Man, tell me about it! :nono:

But it is still my goal to become a professor one day (I'm a postdoctoral fellow now) so that I can tell people what to do even though I would have no clue as to what's going on.

szbd said:

Talking about machine learning, you know sparse data problem? NP complete problem? These always occur in detecting using sonar.

Click to expand...

Of course, these are basic concepts in machine learning. Pfffft.

So tell me, how is learning probabilistic models of signals and then using these models to detect them amongst noise not a prototypical machine learning problem?

bladerunner said:

Hmmm, can you supply us the source of your opinion please:coffee:

Click to expand...

Personal observation. Admittedly that means it is just one data point. However, what is clear is that Chinese engineers are not incapable of mastering the cutting edge data processing tools. The remaining question is if the military research complex has the accumulated wisdom and seasoned senior engineers to direct the efforts of the bright young researchers.

crobato said:

You forgot to mention that sonar travel speed changes with water temperature, density, alkalinity and salinity. Its not a constant like the speed of light, which is with radar. The speed of travel from source to target and back again is what is used with both to determine range. And if speed varies, then so is the accuracy of range determination.

Click to expand...

That's why I mentioned sound in water almost never travel in strait line.