CN106896935A - Input method - Google Patents
Input method Download PDFInfo
- Publication number
- CN106896935A CN106896935A CN201710095765.1A CN201710095765A CN106896935A CN 106896935 A CN106896935 A CN 106896935A CN 201710095765 A CN201710095765 A CN 201710095765A CN 106896935 A CN106896935 A CN 106896935A
- Authority
- CN
- China
- Prior art keywords
- input
- input method
- content
- user
- word
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/02—Input arrangements using manually operated switches, e.g. using keyboards or dials
- G06F3/023—Arrangements for converting discrete items of information into a coded form, e.g. arrangements for interpreting keyboard generated codes as alphanumeric codes, operand codes or instruction codes
- G06F3/0233—Character input methods
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
Abstract
The present invention gives a kind of input method, comprise the following steps, step 1, obtain environment semantic field；Step 2, is input into corresponding input content；Step 3, corresponding output content is given according to environment semantic field and the built-in database of the input method；Step 4, selects the output content for determining.
Description
Technical field
The invention belongs to Computer Natural Language Processing field, relate more specifically to improve defeated by constructing environment semantic field
The method for entering method candidate item precision and realizing prediction type input.
Background technology
In Computer Natural Language Processing field, it is commonly known as " computer semantics recognition " to semantic treatment.It is right
In computer, either semantic model, semantic instance, or semantic applications engine are all by programmer's input computer and consolidate
Change among computer, when user operate, it is according to the historical trace of built-in system accumulation to the input content of user
Judged, provided multiple options, and it is this carry out intellectual analysis according only to historical trace, the accuracy for obtaining result be compared with
Low, it is the problem of urgent need to resolve.
In addition, in computer-aided translation field, applicant once presided over and developed《Refined letter CAT》, the system at present still
Using.As shown in figure 1 and 2, current sentence is split as word and/or phrase sequence to its operation principle by step 1；Step 2,
List each word and/or the corresponding translation of phrase；Step 3, user selects corresponding translation by mouse and shortcut；Step
Rapid 4, it is combined into final translation.
It is this largely once to be generated instead of the software of translator's input service in the input method efficiency very low epoch
Very big effect, certain CAT systems also solve other problemses, such as translation memory, form conversion etc..There is problems in that：
, when a sentence is translated, the ratio that can be fully achieved by mouse click is not high, and input method is also commonly used for translator
To supplement words, this mode toggled between mouse and keyboard can influence operating efficiency.
The content of the invention
In order to solve above-mentioned technical problem, The present invention gives a kind of input method, comprise the following steps, step 1, obtain
Environment semantic field；Step 2, is input into corresponding input content；Step 3, according to environment semantic field and the built-in number of the input method
Corresponding output content is given according to storehouse；Step 4, selects the output content for determining.
Preferably, the environment semantic field includes history input record, currently displaying context on screen, often in screen
The word of upper reading and the document of input method.
Preferably, the built-in database of the input method include China and foreign countries' dictionary database in one or more.
Preferably, the input content is letter.
Preferably, the output content is Chinese.
Preferably, the output content includes individual character, phrase, sentence and/or paragraph.
Preferably, input content described in input through keyboard is used in the step 2.
Preferably, selected using keyboard in the step 4.
Preferably, the input content can be the Chinese phonetic alphabet or five.
The present invention also provides a kind of electronic installation with above-mentioned input method.
Computer Natural Language Processing system of the invention can currently show according on the history input record of user, screen
Context environmental, the frequent all kinds of words read on screen and the active shown are submitted to new defeated based on the present invention
Enter document of method etc., background during collectively as input, establishment is referred to as the database of environment semantic field, and is operated according to user
Historical record, corresponding semanteme is remembered and is sorted, using study and artificial intelligence ability, can be according to user keystroke meter
Corresponding candidate is calculated, and provides order of priority, preferential reference resources are selected out for user, the standard of candidate item can be significantly increased
Exactness simultaneously provides prediction type input.
A kind of simple but typical implementation is translation special input method.Nowadays, all kinds of translation software arise at the historic moment, and
And quite popularize, by taking English-Chinese translation as an example, if user is before the word that one section of translation (block) is displayed on screen is expected,
Can be selected by the character block, be replicated, dragging the methods such as choosing to inform that input method needs the target translated, then can be with
The correspondence Chinese explanation of all words and phrase in selected character block is obtained by inquiring about the method for corresponding English-Chinese dictionary, and
This section of translation of word now can be obtained by calling automatic translation engine.These vocabulary obtained by consulting the dictionary and automatic
The vocabulary included in translation result is that system is based on the usage history of user by study, accumulation and judges and selection, it should
It is content that user most possibly expects input, can be as preferred candidate, this can necessarily improve the accurate of first-selected word
Degree.The Chinese in the semantic environment in whole sentence or whole section, the context between English word or phrase, automatic translation is contacted again
Context and Chinese and English language model between vocabulary and phrase, it is accurate when being input into whole sentence or word fragment
Degree can all bring very big help, in some instances it may even be possible to provide point-device prediction type candidate item, such as, when first Chinese character of key entry
Coding when, system can prompt several most possible word fragments as candidate item, and system can be according to study and product
Tired semantic field is ranked up to each candidate item, in order to select, so as to the operating efficiency of translator greatly improved.
The present invention carrys out constructing environment semantic field to improve candidates of input method in the way of similar to human thinking is imitated
Precision simultaneously realizes that prediction type is input into.
Brief description of the drawings
Fig. 1 is the operating method of existing supplementary translation；
Fig. 2 is the algorithm of existing supplementary translation；
Fig. 3 is the schematic diagram of one embodiment of the present of invention；
Fig. 4 is the schematic diagram of the second embodiment of the present invention；
Fig. 5 is the schematic diagram of the third embodiment of the present invention；
Fig. 6 shows the flow of input method of the invention；
Fig. 7 shows the schematic diagram of fourth embodiment of the invention；
Fig. 8 shows the schematic diagram of fifth embodiment of the invention；
Fig. 9 shows a kind of classified inquiry mode of the invention.
Specific embodiment
It is explained below for implementing embodiments of the invention.It should be pointed out that embodiments discussed below is of the invention
One exemplary embodiment, and those skilled in the art can fit to it without departing from the scope of the gist of the present invention
When modification and similar transformation.Therefore the present invention is not limited by specific embodiment as disclosed below.
For computer system now, it is necessary to the entry for recording user input is easily achieved, and exists now
Most of input method can accomplish this.For on screen show word, can by it is full frame take word technology, OCR technique,
The Automation object that application program is provided, such as, the com object of Microsoft OFFICE, the blind person of operating system reads interface etc.
Technology and obtain.For example, user is when English to Chinese translation is carried out, by obtaining original text, you can by consulting the dictionary and adjusting
Entry and word fragment, the i.e. combinations of words that user's most probable is input into are obtained with automatic translation engine.For another example, user is using
When such as QQ chat, wechat chat instant messaging the inside chat, if we know that the chat content that above both sides are input into,
Just can more accurately provide most probable candidate word.
A kind of simplest implementation method is exactly translation application, such as, user can select, replicate or one section of original of drag and drop
Text, and inform input method module.And most complicated but effective implementation method is exactly to carry out analysing screen according to certain strategy
All words on curtain, all kinds of spoken and written languages that study user often reads, and build corresponding user knowledge base, for
Word, words of frequent input that family is often read etc. may correspond to corresponding private database, and make the database in reality
The algorithm of candidate item generation is participated in during the input operation of border.According to the words that user is commonly used, the database can be carried out automatically
Constantly accumulation and optimize, corresponding informance automatically generated for the input energy of user, and be aware of the specialty of user, interest and
After common words etc., then input method of the invention just has more preferable specific aim and preferential knowledge.Therefore, input method of the invention
It is a kind of more efficient, more intelligent new input method.
Additionally, for further auspicious solution invention, being illustrated by enumerating a very simple example.Such as, when
User is translating in short " Long live to the People ' s Republic of China！", then by consulting English
Chinese dictionary can obtain two phrases：" long live " and " People's Republic of China (PRC) ".And logical available " the Chinese people of automatic translation engine
Republic's long live！”.In addition, working as user input：Zhonghua (China) or zh (brevity code of " China "), or even only it is input into z
When, input method of the invention may intelligently provide the candidate item of prediction：" 1, long live the People's Republic of China！" and " 2, in
Magnificent people's republic " etc., and accumulated experience according to the options of user, the candidate item predicted is entered according to selected probability
Row sequence., merely with being illustrated as a example by Pinyin coding, other coded systems are similarly for the present embodiment.
Although of the invention improve candidates of input method precision and realize what prediction type was input into by constructing environment semantic field
Also there are some aspects to be modified in method, but the meaning of a word of wherein most or word fragment are correct.For its use
On the way, due to the prompting and filtration of user keystroke, can easily win useful part therein and be used, therefore,
There is wide application prospect in the method for the present invention.
The present invention can be significantly increased the accuracy of the candidate word of input method by constructing environment semantic field, reduce
Keystroke, and can develop that degree of intelligence is higher, can precisely predict the input method of user view.
Prediction input of the invention is sets forth in Fig. 3-Fig. 5, the embodiment with environment candidate is rapidly input.
The first embodiment of the present invention is shown in Fig. 3, a kind of situation of prediction input shows such as figure institute on screen
The context 112 shown, now input method of the invention by the context have been carried out collection, when user input
During input content " ZH " 112, except the conventional database 113 of display (can be various dictionaries or those skilled in the art
The various dictionaries known) in, while predicting candidate 114 is also show, as long as the corresponding output content 113 of user's selection,
114, it is possible to expeditiously complete preferably operation.
(rapidly input) figure 4 illustrates the second embodiment of the present invention, input method is by looking into word in this embodiment
Allusion quotation provides environment candidate item 114', while ordinary candidate 113' is given, wherein showing context fragment on screen
111'。
Figure 5 illustrates the third embodiment of the present invention (looking into word input), wherein have input input content 112 ", and
Here, input method of the invention can be according to the context (environment word) 111 on screen " reduce user quantity, i.e., only need
The fragment of English word or word is input into, option one 14 now can be just obtained under this input method ".
Fig. 6 shows algorithm steps of the invention, in step 11, obtains environment semantic field；Step 12, input is corresponding defeated
Enter content；Step 13, corresponding output content is given according to environment semantic field and the built-in database of the input method；Step 14,
Select the output content for determining.
Fig. 7 and Fig. 9 show other situations of the invention.
Above by taking supplementary translation as an example, the input method specific function to be had of " there is eyesight " is illustrated.As
Basis, can stretch to deeper, more extension, produce all new generation input method.
Common input accuracy is substantially improved
Due to the introducing of context environmental, candidate can be calculated with more complicated language model, time is necessarily substantially improved
Precision, more traditional input method is selected to have more than 30% increased quality.
As shown in fig. 7, user is currently in use TALKER, then due to it is recognised that nearest chat content
111 " ', it is of course possible to the idea of user is more accurately guessed, so as to provide more accurate candidate or even answer suggestion.Such as, other side
A New cinema just is mentioned, and you will mention the name of this film in answer, for our new input method, use
Family can treat this " neologisms " as old friend completely, can simply enter 112 as well known vocabulary " ', and
Shorthand way can be used！Such as this film noun then works as user input " Ah according to buy ox Grand Canyon " " a.g ", i.e. fragment
The initial of one and last Chinese character, centre plus a fullstop represent omissions, output content 114 " ', " Ah is according to buying Niu great Xia
Paddy " just appears in candidate first place at once, very convenient.
Simultaneously
If at this moment pointing out the movie related information 115 on " Ah according to buy ox Grand Canyon " side, those skilled in the art hold
Be readily conceivable that, such relevant information 115, such as brief introduction, act the leading role ... will cause chat it is more interesting, and also can directly with purchase
Bill system is connected
In addition, people are at writing (summary is all to write article, composition, novel, sheet of language), mainly by brain
Design, is expressed, and generally require some intellectual contents by keyboard, can not thought for the moment or is not remembered clearly, not sure, it is possible to
Seek help other application, such as browser goes to search.At this time not only hand have left keyboard, in omnifarious webpage
It is easy to lose, or even has forgotten original intention, design can be also interrupted.
Figure 8 illustrates a kind of situation, want to quote the stich of li po, but only remember wherein fragment 111 " ", such as one
Individual word, " Chu Jiang ", then be input into 112 in input method " ", that is, obtained output result 114 " " be " li po Chu Jiang " Chinese, i.e.,
Can obtain full poem 115
Further embodiment of the invention is shown in Fig. 9, desired result is inquired by keying in corresponding contents.
Since input method can do many things, then also can just be sent out to intelligent assistant direction using it as control centre
Exhibition.Some tasks are such as completed, is collected and is arranged certain class data, some work are automatically performed as real assistant.It is such
No matter new input method with traditional input method compared with, all will occur qualitative change from function or degree of intelligence, its application prospect without
Limit！By taking chat scenario as an example, the entirely possible question recommending by other side is several to answer as candidate, as long as user selects！
Although below set forth principle of the invention in conjunction with specific embodiments it should be appreciated that the present invention
Disclosed exemplary embodiment is not limited to, the scope of following claims is limited to wrap according to its broadest scope
Include all modifications.
The various embodiments described above are only the preferred embodiment of the present invention, in the art, every based on skill of the present invention
Changes and improvements in art scheme, should not exclude outside protection scope of the present invention.
Claims (10)
1. a kind of input method, comprises the following steps, step 1, obtains environment semantic field；Step 2, is input into corresponding input content；
Step 3, corresponding output content is given according to environment semantic field and the built-in database of the input method；Step 4, what selection determined
Output content.
2. input method according to claim 1, it is characterised in that the environment semantic field includes history input record, screen
Currently displaying context on curtain, the word and the document of input method often read on screen.
3. input method according to claim 2, it is characterised in that the built-in database of the input method includes China and foreign countries' dictionary
One or more in database.
4. input method according to claim 3, it is characterised in that the input content is letter.
5. input method according to claim 4, it is characterised in that the output content is Chinese, English or other mesh
Poster is sayed.
6. input method according to claim 5, it is characterised in that the output content includes individual character, phrase, sentence and/
Or paragraph.
7. input method according to claim 6, it is characterised in that using in input described in input through keyboard in the step 2
Hold.
8. input method according to claim 7, it is characterised in that selected using keyboard in the step 4.
9. input method according to claim 8, it is characterised in that the input content can be the Chinese phonetic alphabet or five
Pen.
10. there is the electronic installation of the input method according to any one of claim 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710095765.1A CN106896935A (en) | 2017-02-22 | 2017-02-22 | Input method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710095765.1A CN106896935A (en) | 2017-02-22 | 2017-02-22 | Input method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106896935A true CN106896935A (en) | 2017-06-27 |
Family
ID=59185038
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710095765.1A Pending CN106896935A (en) | 2017-02-22 | 2017-02-22 | Input method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106896935A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109656385A (en) * | 2018-12-28 | 2019-04-19 | 北京金山安全软件有限公司 | Input prediction method and device based on knowledge graph and electronic equipment |
| CN113625885A (en) * | 2020-05-08 | 2021-11-09 | 北京搜狗科技发展有限公司 | Input method, input device and input device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102591472A (en) * | 2011-01-13 | 2012-07-18 | 新浪网技术(中国)有限公司 | Method and device for inputting Chinese characters |
| CN106125955A (en) * | 2016-06-23 | 2016-11-16 | 百度在线网络技术（北京）有限公司 | A kind of method and apparatus that hot word is provided in applying in input method |
| CN106293119A (en) * | 2016-07-29 | 2017-01-04 | 百度在线网络技术（北京）有限公司 | A kind of method and apparatus carrying out information recommendation in input method |
| CN106371624A (en) * | 2016-09-23 | 2017-02-01 | 百度在线网络技术（北京）有限公司 | Method and device for providing input candidate item |
-
2017
- 2017-02-22 CN CN201710095765.1A patent/CN106896935A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102591472A (en) * | 2011-01-13 | 2012-07-18 | 新浪网技术(中国)有限公司 | Method and device for inputting Chinese characters |
| CN106125955A (en) * | 2016-06-23 | 2016-11-16 | 百度在线网络技术（北京）有限公司 | A kind of method and apparatus that hot word is provided in applying in input method |
| CN106293119A (en) * | 2016-07-29 | 2017-01-04 | 百度在线网络技术（北京）有限公司 | A kind of method and apparatus carrying out information recommendation in input method |
| CN106371624A (en) * | 2016-09-23 | 2017-02-01 | 百度在线网络技术（北京）有限公司 | Method and device for providing input candidate item |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109656385A (en) * | 2018-12-28 | 2019-04-19 | 北京金山安全软件有限公司 | Input prediction method and device based on knowledge graph and electronic equipment |
| CN113625885A (en) * | 2020-05-08 | 2021-11-09 | 北京搜狗科技发展有限公司 | Input method, input device and input device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA3129745C (en) | Neural network system for text classification | |
| Khan et al. | A novel natural language processing (NLP)–based machine translation model for English to Pakistan sign language translation | |
| US8903707B2 (en) | Predicting pronouns of dropped pronoun style languages for natural language translation | |
| Fetaya et al. | Restoration of fragmentary Babylonian texts using recurrent neural networks | |
| Kumaran et al. | Compositional machine transliteration | |
| CN108932218A (en) | A kind of example extended method, device, equipment and medium | |
| Jain et al. | “UTTAM” An Efficient Spelling Correction System for Hindi Language Based on Supervised Learning | |
| Toselli et al. | Transcribing a 17th-century botanical manuscript: Longitudinal evaluation of document layout detection and interactive transcription | |
| CN108009158A (en) | Interaction prompts interpretation method, device, storage medium and electronic equipment | |
| Farooq et al. | A multi-stack RNN-based neural machine translation model for English to Pakistan sign language translation | |
| Vandeghinste et al. | Improving the translation environment for professional translators | |
| Haq et al. | Urdu named entity recognition system using deep learning approaches | |
| CN106896935A (en) | Input method | |
| Kennington et al. | Incrementally tracking reference in human/human dialogue using linguistic and extra-linguistic information | |
| Ruma et al. | Transformer based answer-aware bengali question generation | |
| Neumann et al. | Corpus methodology and design | |
| Che et al. | A word segmentation method of ancient Chinese based on word alignment | |
| Laviosa et al. | A corpus analysis of translation of environmental news on BBC China | |
| Dombrowski et al. | Language Is Not a Default Setting: Countering DH’s English Problem | |
| US9311302B2 (en) | Method, system and medium for character conversion between different regional versions of a language especially between simplified chinese and traditional chinese | |
| Zhang et al. | Icon metaphor design based on ideographic Chinese character-creation | |
| Xu | Digital Approaches to Understanding Dongba Pictographs | |
| Domingo et al. | A comparison of character-based neural machine translations techniques applied to spelling normalization | |
| Li et al. | Interpersonal Interface System of Multimedia Intelligent English Translation Based on Deep Learning | |
| US20230394250A1 (en) | Method and system for cross-lingual adaptation using disentangled syntax and shared conceptual latent space |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170627 |