What I want, basically is a natural language. Just this direction, I need to learn more. At present, I am watching some professional books; but I am even more interesting to connect the previous lens. I feel very interesting. I. I also tested a short message of the word, is the first paragraph of an article, very obvious personal style. as follows:
Reading 10,000 books, Wanli Road, is the two wishes in this life. Reading, it is a legacy; the university has been too half, one time, in addition to the teacher's joy, East Lake Boss, it is the book of the book. However, the computer is professional, and the book is quite spent. So get a reference to the partner of the unpacking, column, and give a reference; however, I don't have advertisements ~ I will explain the following:
The results of massive quota demo are as follows:
Reading 10,000 books, Wanli Road, is the two wishes in this life. Reading, it is a old habit; the university has been too half, one time, in addition to the teacher's joy, Donghu Biwa, is the book of the book. However, the computer is professional, and the book is also quite spent. So get a reference to the column of column, column, give a reference partner; however, I am not advertising ~ I have said as follows:
Most of them are still nice. But watching "One Town", this is the most critical, because the ambiguity of "back" is the largest, you can be the word before and after; "I mean here is of course" back ", you can regret. Although the word is a difficult point, I don't think it can be seen separately to emphasize the word. In the same paragraph, I use Unicom International Instant Translation to English, the effect is as follows:
It is two hopes of this life to read ten thousand books and walk ten thousand li Study and can be regarded as old to practise;. University already too much halves, turn one's head at one, at teachers and friends joyous faces, East lake and bluish waves are books of that volume But the speciality of the computer, does not win the others than, the book is quite a piece of cost It tidies up it arranges pieces of list give silly the one of partners of book consult;.. However , I am NOT Advertising But Must Prove As Follows: Time: 10.428 Sc
The judgment of the language and sentence is still a bad thing; but from the Turn One's Head At One, "one time" is still pair :). If you use the same server to translate this English into a Chinese, God! I don't know! as follows:
Two hopes to read 10,000 books and walk the study of the life of 10,000 milligramists and may be treated as the old practice; have been universities more than half a half, turn a man in one, in the teacher and friends , Full of happy faces, east lake and blue waves are books. But the professionalism, ratio, book of the computer is quite good, don't win other. The discussion of the discussion of the debris directory to the partner of the stupid book; however, I am not advertising, but I must prove that as follows: this processing result, let me think this road is still growing. Do you think? In addition, I am a large number of students, and there is still a lot of science and technology. I want to learn more about the knowledge of natural language, can the company have introduced several books? thank you very much! Second, about HNC's counter comment 1, I also read the book of HNC theory, but I don't think I don't agree with "concept symbol hypothesis". In fact, I can't see HNC with Jumsky's generators and their disciples evolved. What is the difference between the generation semantics. It is basically unob possible to expect HNC to solve the problem of natural language understanding, especially for oriental languages. Another point is that HNC and the Jumsky theory therezetted have not considered the evolution of language acquisition and language. From the perspective of language changes, HNC can't stand because it is impossible to use symbolic intelligence methods to solve the self-study of the perfect concept mapping. So I think it is still based on the method of connectionism based on corpus, I can finalize NLU issues. It is unfortunate that the HMM method does not perform a well-behaved structure, which stops the word library method. Xcool comment 2, HNC is just a kind of understanding, walking is based on rule-based NLP old road, and this road has proven that the large-scale real text cannot be processed. Mr. Huang is characterized by not concerning the mainstream technology of international and domestic language processing, and I have a disciple to make a mysterious thing. Due to its theoretical defects, HNC can't make things that can be used for Demo, even if you have a product or available to Demo, you can also achieve better performance. Please be careful. CHRIS Reviews 1, there is a reference to the following reference, Jincao Red Mr. Jin and I have answered by email: "About netizens's views, it is indeed a view of many people. But HNC and Jumsky's generatory speech method is There is different, the latter is concerned about the generation of language, and the former is concerned about the language understanding, understanding and generation of the mapping and reflection of the natural language to language concept space, and the focus of understanding is semantic correlation ( Is it possible to understand it correctly), the focus of the generation is the normative syntax (whether the sentence generated is syntax). "
"About" The Evolution of Language Acquisition and Language "is the latter language of the language. It is the problem of language scholars care, and our current focus is to let the computer reach a primary school student, the level of middle school students, I still don't expect this level. You If you ask a primary school, what is "the evolution of the language", I am afraid you can't get the answer. Similarly, you should not ask your computer to do it. "" The above view is for discussion. "As for comment 2, no Talk. Now there are quite a few theoretical parts being started to be created. The corresponding productization work is also started. The theory itself has a lot of defects, of course, has to be inspected, and it has to be continuously improved. I really hope that the HNC team can make persistent efforts to go forward.
