Announcement

Collapse
No announcement yet.

Intel Releases ControlFlag 1.0 For AI-Driven Detection Of Bugs Within C Code

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Intel Releases ControlFlag 1.0 For AI-Driven Detection Of Bugs Within C Code

    Phoronix: Intel Releases ControlFlag 1.0 For AI-Driven Detection Of Bugs Within C Code

    Intel last month open-sourced "ControlFlag" for finding bugs within source code by using AI with training off more than a reported one billion lines of code. Intel has said they have successfully been using it within their software from applications down to firmware. The new milestone today is ControlFlag 1.0 being released...

    https://www.phoronix.com/scan.php?pa...ontrolFlag-1.0

  • #2
    That's pretty cool. I'd imagine machine learning would be great at this sort of processing huge amounts of data and finding anomalies.

    Comment


    • #3
      ... And next comes the GPLv4, denying the use of source code for the purpose of training an AI unless the AI itself is open-sourced under a GPL license.

      Luckily is it under the MIT License.

      Comment


      • #4
        So humans have to define what is "good code".

        Hopefully the definition will be "does exactly what it needs to, doesn't do anything it doesn't need to, with zero bugs, unexpected behaviours or vulnerabilities".

        How much code fits that definition?

        print("Hello world!")

        And even that is doubtful depending on the language chosen.

        Comment


        • #5
          Originally posted by kvuj View Post
          That's pretty cool. I'd imagine machine learning would be great at this sort of processing huge amounts of data and finding anomalies.
          No such tool is perfect. But perfect is the enemy of good enough, and the various code analysis tools have identified code areas that are, at least, sufficiently questionable to be reviewed. I will take and review such advice each and every time.

          Comment


          • #6
            Originally posted by Paradigm Shifter View Post
            So humans have to define what is "good code".
            I'd start with Spark ADA as a requirement and then work from there ...

            Comment


            • #7
              Originally posted by sdack View Post
              ... And next comes the GPLv4, denying the use of source code for the purpose of training an AI unless the AI itself is open-sourced under a GPL license.

              Luckily is it under the MIT License.
              there are ai driven snippet generators, but the question arises: what license does the generated code follow? It could stem from gpl.

              Comment


              • #8
                Originally posted by oleid View Post

                there are ai driven snippet generators, but the question arises: what license does the generated code follow? It could stem from gpl.
                Well, replace 'AI' with 'human', and copyright law applies, with the attendant problems of defining what is a 'fair use' sized snippet. My view would be is that if any GPL code were used to train the AI, then either all the AI's output must be GPL licenced or generated code should be accompanied by a decision tree record (or equivalent) showing all generated code was derived from sources with licences compatible with the licence proposed for the code output by the AI.

                The problem is quite simple: don't train the AI on GPL licensed code unless you are happy for the AI's output to be GPL licensed. The same would apply for BSD-n licensed code, or indeed any other licence other than 'public domain'. Essentially, AIs should be trained on public domain code (if it exists) together with code licensed with licences compatible with the licence you expect the AIs code to be used under.

                Comment


                • #9
                  Originally posted by Old Grouch View Post

                  Well, replace 'AI' with 'human', and copyright law applies, with the attendant problems of defining what is a 'fair use' sized snippet. My view would be is that if any GPL code were used to train the AI, then either all the AI's output must be GPL licenced or generated code should be accompanied by a decision tree record (or equivalent) showing all generated code was derived from sources with licences compatible with the licence proposed for the code output by the AI.

                  The problem is quite simple: don't train the AI on GPL licensed code unless you are happy for the AI's output to be GPL licensed. The same would apply for BSD-n licensed code, or indeed any other licence other than 'public domain'. Essentially, AIs should be trained on public domain code (if it exists) together with code licensed with licences compatible with the licence you expect the AIs code to be used under.
                  That makes so much sense it'll never work.

                  I'm getting back to the eclipse. 31F so I'm freezing my ass off.

                  Comment


                  • #10
                    Originally posted by Old Grouch View Post
                    The problem is quite simple: don't train the AI on GPL licensed code unless you are happy for the AI's output to be GPL licensed. The same would apply for BSD-n licensed code, or indeed any other licence other than 'public domain'. Essentially, AIs should be trained on public domain code (if it exists) together with code licensed with licences compatible with the licence you expect the AIs code to be used under.
                    That is the obvious answer I think. Those who disagree are jumping through hoops as they want the best of both worlds (freely usable output code, trained on any license).

                    Note that 'public domain' isn't the only answer. Microsoft could have trained it on their proprietary code too. Did they? I'm not sure they did, for "obvious" reasons. Double standards... It can be proprietary code if the license allows training with it.

                    Now, of course MIT and BSD can in most cases be combined with proprietary code... Github should have restricted their training set.

                    Comment

                    Working...
                    X