In my previous blog "Can Machine Understand What You mean?", Qualrus was introduced to be used to grade student's assignments, in particular essay. I could not agree this brave experiment. The reason is to my current knowledge computers don't hold the abilities to give out a judgement for the rightness or wrong. As for the personal subjective ideas in the essay, it is impossbible to be graded by machines.
Similiar news about Brent's software makes me believe with no doubt that computers could not replace humans/teachers to do essay grading work. It is a big challenge, but will not succeed under the currrent circumstances due to the machine's lack of re-birthing and evolveing of new concepts from old concepts, and connecting between relevant relationships to compose reasoning process naturally (to my sense). The huge backend behind this grading software must be a huge knowledge base, or simply a huge database, which provides information sources (example sentences, already-scored papers) for the comparison work between students' essays submitted to the system and sample essays in the backend source then give out a score finally based on those similiar-quality examples. There are different ways to express the same idear! How we could guarantee the consistency and integrality of the sources will be unresolvable with our current technologies!
To the end, computers are still machines! They are operated by the human beings. Actually the origins of this kind of software/systems, I think, are coming from the people's desires to seek the equity. They do not hope that "different teachers grade different papers differently". I really don't want to know that the real answers for this kind of software is to free human beings from responsibilities. That will make us learn much faster how to cheat computers than human beings. It is the tragedy of advanced technologies!
-
No comments:
Post a Comment