標題: Bias This problem is very similar to the first [打印本頁] 作者: [email protected]時間: 2024-3-12 11:59 標題: Bias This problem is very similar to the first It is also used in hundreds of modern military uses and it even has a major role in machine translation and improving communication between humans and each other. CHALLENGES AND PROBLEMS FACING DEEP LEARNING Despite all the great things we mentioned about Deep Learning there are also a large set of challenges and problems facing it most notably Explanation and slides of their results Explainability of Outputs They call this problem the Black Box.
Unfortunately you do not know how the deep intelligence model reached these results which may cause a lot of problems in some cases if the model was used. In a different way than you expect imagine for example that it evaluates wBelgium WhatsApp Number Datahether or not you are likely to deserve your loan from the bank based on the size of your nose. in that the model works on the data that you input to it so for example if the model calculates the probability of you committing a murder from your photo a theoretical model only for clarification and you feed it with photos of black people only It is possible for every black person to be classified as likely to commit murder.
The enormous energy required to run the models Smart models that need all this intelligence and analyzing all that data consume lots and lots of computing power. Privacy and Data Protection There are many fears and problems about the possibility of security problems and hacks that expose sensitive data to danger. Its impact on humans Deep learning may have a very serious impact on humans and I just have to remind you that it can change the reality of work for the worse on several levels and not just by eliminating a lot of jobs and making a lot of people unemployed.
