Many students think that the minimum detection unit of HowNet duplicate checking is sentence. As long as every sentence is not repeated.

But in fact, HowNet system adopts semantic level detection technology. The system will combine the content of the context to judge the content that reaches a certain semantic level, not just based on one sentence.

In vernacular Chinese, we can't just adjust the word order of references. For example, we can muddle through by changing "Zhai copied other people's papers" to "others' papers were copied by Zhai". Not to mention the whole paragraph plagiarism, only the order of paragraphs and sentences was adjusted.

Second, refer to common network literature or resources.

Many students like to find some reference materials or documents through online search when writing papers. However, the HowNet duplicate checking system will also conduct online search when detecting articles. In particular, some well-known websites with a lot of online academic materials are the key areas of HowNet, and whoever uses them will check it.

So how to avoid it?

In fact, we can try to search foreign literature and translate it into Chinese. Because everyone's language habits and English level are different, even the same words will be different after translation.

Third, it is impossible to search by consulting related books.

Many students are smart enough to think that it is impossible for HowNet to find those early professional books, and there is no electronic version. I don't know how many people have cited this book in their papers before, and it has been archived in the discussion library and online. Therefore, when quoting books, it is best to use online translation tools to translate them into small languages and then translate them back into Chinese. Then change the word order to fluency.

I would like to remind you not to use common search engines, because many of your previous papers have used these common online translations to steal columns, so it is safer to find other online translation websites.

Fourth, do not pay attention to the format of references.

As we all know, HowNet will automatically identify references and will not conduct text detection. However, many students' references will still be marked as duplicate texts in red after duplicate checking. This is because the format of the search literature is incorrect and is treated as text by HowNet system.

Knocking on the blackboard: reference documents with completely standardized format should be displayed in gray font in the inspection report of HowNet. Any reference displayed in red, yellow or orange is due to a format error.

Fifth, picture plagiarism can't be found on HowNet.

Stealing copyrighted original pictures, experimental image data, charts, etc. Being from others is also plagiarism. But it's not that many friends don't realize it. However, it is generally recognized that HowNet does not have the ability to scan and identify the engine in the picture.

Including the strategy of duplicate checking of many papers on the Internet, all of which teach you how to turn words into pictures and insert them into papers to avoid the detection of HowNet. This is no problem at present, but in today's increasingly mature OCR technology, the author estimates that HowNet will look up the map again soon. So please don't take any chances when writing your graduation thesis.

Sixth, you don't need to predict your handwritten paper before checking it again on HowNet.

This is also a low-level mistake made by many students. I think as long as it is a paper written by myself, there is definitely no problem in duplicate checking. I don't know how strict the detection system of HowNet is now, even a purely handwritten paper will "crash" with other papers on the Internet by mistake. If the school detection rate is strict, it is likely to be a false and wrong case.