zunia.vn

Tuyển sinh 2024 dành cho Gen-Z

zunia.vn

» Luận Văn - Báo Cáo

» Báo cáo khoa học

Báo cáo khoa học: "Mechanical Translation Work at the University of Michigan"

Chia sẻ: Nghetay_1 Nghetay_1 | Ngày: | Loại File: PDF | Số trang:0

Báo xấu

57
lượt xem 2
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

THE PRINCIPAL differences between the work at The University of Michigan and other work in machine translation is in the emphasis placed on the problem of multiple meaning and the approach to that problem. Our approach consists in translating small groups of words, listing in the dictionary multiple meanings under each word in the group, and finding algorithms which make it possible to choose the proper set of meanings for the group.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Báo cáo khoa học: "Mechanical Translation Work at the University of Michigan"

[Mechanical Translation, vol.3, no.2, November 1956; pp. 34, 41] Mechanical Translation Work at the University of Michigan A. Koutsoudas and R. Machol, Willow Run Laboratories, University of Michigan of type for technical periodical literature. It is THE PRINCIPAL differences between the work assumed that within a generation machine trans- at The University of Michigan and other work lation will be a fait accompli, as will machine in machine translation is in the emphasis placed reading (i.e., the scanning of printed matter on the problem of multiple meaning and the ap- with the production of signals suitable for proach to that problem. Our approach consists feeding a computer). All of the great mass of in translating small groups of words, listing in technical periodical literature will then be rou- the dictionary multiple meanings under each tinely translated into many languages. At that word in the group, and finding algorithms which time a number of trivial problems will arise, make it possible to choose the proper set of involving differences in type faces (fonts), dia- meanings for the group. Some 9f the dictionary critical marks, displayed matter (e.g. equations), meanings under each multiple-meaning word underlining, the use of italics or boldface to will be vacuous and some will be redundant. convey special meaning, etc. The algorithms are based on the pattern of va- When mechanical reading and translation are cuous translations in the dictionary for the routine, these trivial problems will be solved group of words under consideration. For ex- by international standardization. However, ample, for a particular idiomatic three-word this will leave the great bulk of the technical sequence, the fourth meaning under the first and literature published in the intervening years third words might be vacuous, and the entire either untranslatable or translatable only with idiom will be translated under the second word. great extra difficulty. It is therefore suggested The algorithm will be such as to lead the ma- that this standardization be performed now, so chine to pick the fourth meaning for each word that all technical literature published after, say in this case. These algorithms are discussed 1960, would be translatable by machine. As a in more detail in the article on page first step it is suggested that a universal font Since the problem of multiple meaning cannot be established. For this purpose it will be ne- be solved apart from the entire problem of cessary to make the following studies: (1) The t ranslation, rules are also being prepared for readability of various fonts, from the human the syntactical and grammatical aspects of trans- engineering point of view (accuracy and speed) lating Russian into English, and a large corpus and from the publisher's point of view (appear- of Russian is being processed. At the present ance and reader satisfaction). (2) The machine time 64,000 running words (128 pages) of mate- requirements. This will involve some crystal- rial from the Journal of Experimental and The- ball estimates as to what the finally successful oretical Physics is being coded onto punched reading device will be like. Of course, such cards, and experiments are being carried on in machines will eventually be able to cope with which technicians simulate a computer in trans - certain differences, but their task will be made lating according to the stated rules. Theoretical enormously easier if they do not have to cope frequency studies are also underway. These with the difference between K and K or between studies will use the results of the punched-card T a nd T . analysis. The theoretical aspects are based on It may be possible to standardize also on equations comparable to those of Zipf's law. It certain other things. For example, most equa- is hoped to be able to predict answers to such tions are numbered, in parentheses, at either questions as: How many different words will be the beginning or end of the line. It might be found in a million running words? How many possible to standardize on the beginning of the new words will be found in a second sample line, and to use the open-parenthesis sign, (, at equally large? How many words must there be the left to indicate any displayed matter. This in a dictionary to ensure having 99% of the words could be a cue to the machine to photograph in a sample randomly chosen from a certain rather than translate. field ? The University of Michigan also presented to C ontinued on page 41 the meeting a recent idea for a Universal Font
41 KOUTSOUDAS from page 34 Application to non-Roman-alphabet languages posals for international standardization, since (especially Russian) would be a possibility for these people are most likely to cooperate on the more distant future. such matters. Furthermore, the change will After a suitable standard font has been chosen, probably not involve any expense, since the it will be necessary to convince the publishers printers of these journals have hundreds of fonts of technical journals to use it. This should not already and can continue to use the discarded present nearly so much difficulty as many pro- fonts for non-technical publications.

CÓ THỂ BẠN MUỐN DOWNLOAD

LV.09: Bộ Luận Văn Tốt Nghiệp Chuyên Ngành Quản Trị Kinh Doanh

81 tài liệu

1637 lượt tải

THÔNG TIN

TRỢ GIÚP

HỖ TRỢ KHÁCH HÀNG

Theo dõi chúng tôi

Chịu trách nhiệm nội dung:

Nguyễn Công Hà - Giám đốc Công ty TNHH TÀI LIỆU TRỰC TUYẾN VI NA

LIÊN HỆ

Địa chỉ: P402, 54A Nơ Trang Long, Phường 14, Q.Bình Thạnh, TP.HCM

Hotline: 093 303 0098

Email: support@tailieu.vn

Giấy phép Mạng Xã Hội số: 670/GP-BTTTT cấp ngày 30/11/2015 Copyright © 2022-2032 TaiLieu.VN. All rights reserved.