4 ms·

Actually I am against machine learning. It only makes the debugging more difficult. There should be simple rules and not some classifier one cannot understand w

Actually I am against machine learning. It only makes the debugging more difficult. There should be simple rules and not some classifier one cannot understand without a math degree.

lawrenceyan 9y ago
You make it sound like machine learning is some kind of black box. Machine learning is a huge field and there are many analytical methods within it that can offer direct understandable insights into a process.
- SilasX 9y ago
  Rescinding downvote to give an explanation. Yes, ML isn't magic, but that's irrelevant to the parent's point: the issue is that websites a stable "browser API" to work against so they can know what the user will send back where and when. If the user is selectively setting cookies based on the results of machine-learning some non-obvious then it becomes that much harder to debug "why didn't this part of the site remember who they were?"
  - lawrenceyan 9y ago
    But that's what I mean. Machine learning as a field is enormous. There are wide variety of implementations that can be 100% derived with every step of logic being open for people to analyze. It would be no different than building any other form of conditional system.
    - SilasX 9y ago
      That's still missing the point: from the fact that your model derivation is intelligible, it doesn't follow that websites will know that you're using the model, or will be able to product consistent, desirable behavior for both you and everyone else using some different behavior. Sites will break without warning for users until some engineer discovers that some ML model told the browser to only set cookies for a particular set of sites determined by IP, domain, "trustworthiness", and a few other factors engineers don't want to think about when designing the app. If you're saying that you can publish the model openly with the expectation that sites will know you're using it, that's fine, and that's a valid approach, but that has nothing to do with ML per se, and is just another attempt at rearchitecting the browser/server interface with all of its associated issues, and continues the same arms race between sites that want to de-anonymize you. In short: there's a tradeoff between "hiding information about yourself" and "providing a stable set of expectations for web sites to build off of"; ML can favor one of those objectives but not eliminate the tradeoff and somehow get the best of both worlds. The only way to get the best of both worlds is to rearchitect the API by which browser clients talk to webservers so that it's easy to separate out what you do want/need to tell the server vs what you don't.
      - lawrenceyan 9y ago
        That's definitely a valid consideration. I was more focusing on refuting OP's statement in implying machine learning is a black box because only a select few people are able to understand it due its complexity. I think I should have made that more clear in my original comment in order to avoid any confusion.

Actually I am against machine learning. It only makes the debugging more difficult. There should be simple rules and not some classifier one cannot understand without a math degree.

lawrenceyan 9y ago

You make it sound like machine learning is some kind of black box. Machine learning is a huge field and there are many analytical methods within it that can offer direct understandable insights into a process.

SilasX 9y ago
Rescinding downvote to give an explanation. Yes, ML isn't magic, but that's irrelevant to the parent's point: the issue is that websites a stable "browser API" to work against so they can know what the user will send back where and when. If the user is selectively setting cookies based on the results of machine-learning some non-obvious then it becomes that much harder to debug "why didn't this part of the site remember who they were?"
- lawrenceyan 9y ago
  But that's what I mean. Machine learning as a field is enormous. There are wide variety of implementations that can be 100% derived with every step of logic being open for people to analyze. It would be no different than building any other form of conditional system.
  - SilasX 9y ago
    That's still missing the point: from the fact that your model derivation is intelligible, it doesn't follow that websites will know that you're using the model, or will be able to product consistent, desirable behavior for both you and everyone else using some different behavior. Sites will break without warning for users until some engineer discovers that some ML model told the browser to only set cookies for a particular set of sites determined by IP, domain, "trustworthiness", and a few other factors engineers don't want to think about when designing the app. If you're saying that you can publish the model openly with the expectation that sites will know you're using it, that's fine, and that's a valid approach, but that has nothing to do with ML per se, and is just another attempt at rearchitecting the browser/server interface with all of its associated issues, and continues the same arms race between sites that want to de-anonymize you. In short: there's a tradeoff between "hiding information about yourself" and "providing a stable set of expectations for web sites to build off of"; ML can favor one of those objectives but not eliminate the tradeoff and somehow get the best of both worlds. The only way to get the best of both worlds is to rearchitect the API by which browser clients talk to webservers so that it's easy to separate out what you do want/need to tell the server vs what you don't.
    - lawrenceyan 9y ago
      That's definitely a valid consideration. I was more focusing on refuting OP's statement in implying machine learning is a black box because only a select few people are able to understand it due its complexity. I think I should have made that more clear in my original comment in order to avoid any confusion.