CN111652703A

CN111652703A - Method and system for automatic accounting and tax declaration of artificial intelligence accounting

Info

Publication number: CN111652703A
Application number: CN202010499723.6A
Authority: CN
Inventors: 黄云峰; 曹武; 陈志军; 唐丽娟; 冯潇; 刘杨; 刘子超; 芮体江; 芮麟
Original assignee: Cela Artificial Intelligence Technology Yunnan Co ltd
Current assignee: Cela Holdings Yunnan Co ltd
Priority date: 2020-06-04
Filing date: 2020-06-04
Publication date: 2020-09-11
Anticipated expiration: 2040-06-04
Also published as: CN111652703B

Abstract

The invention provides a method and a system for automatic accounting and tax return of artificial intelligent accounting, which are used for realizing the method and the system for automatic accounting and tax return by interaction with a big data center and a neural network through a wireless network or a 4G and 5G network in a smart phone, a portable computer and computer equipment, wherein invoices, bank running water and receipt are shot through a mobile phone camera or automatically acquired through a third-party electronic invoice platform SDK and then transmitted to an application layer through the wireless network or the 4G and 5G network, intelligent accounting services such as accounting vouchers, detailed accounts and expense reimbursement notes are automatically generated for users based on the small enterprise accounting criteria, the enterprise accounting criteria and the tax law as basic algorithm logic, and tax return is automatically executed in a tax return month; the invention greatly reduces the working threshold of accountants and improves the user experience, and the big data center system and the neural network help the user to realize zero-basis financial bookkeeping through deep learning and training iteration of a large amount of accounting bookkeeping behaviors.

Description

Method and system for automatic accounting and tax declaration of artificial intelligence accounting

Technical Field

The invention relates to the technical field of bookkeeping, in particular to an artificial intelligence accounting automatic bookkeeping and tax declaring method and system.

Background

Financial accounting is a highly skilled task that must be matched by the financial staff to the appropriate lending subject for each original document to record the economic activity of the enterprise. In the current financial software, when accounting and operation are performed by financial staff, the mode is as follows: the original voucher (such as invoice, bank statement, reimbursement bill and the like) needs to be compiled into a accounting voucher meeting the financial requirement according to certain financial rules. This presents the following challenges to the practitioner:

(1) a large number of subjects need to be remembered: along with diversification of enterprise business, the structure of financial subjects becomes more complex, units, stocks, raw materials and the like can reach hundreds of items, and when one business occurs, financial staff need to search for a proper subject from hundreds of subjects, which is a time-consuming and labor-consuming work;

(2) very specialized industry knowledge is required: when special industries such as chemical industry, medicine, planting and the like are involved, the original documents can be correctly classified by accounting when inventory, raw materials and biological assets are processed by accounting and a large amount of industry knowledge is required, so that the learning cost is very high;

(3) lack of certain criteria: newly entered accounting is basically taught by old accounting and training institutions from cashiers, and no practical standard is provided to guide accounting to operate more correctly in some cases.

With the continuous development and expansion of internet technology and market scale, more and more enterprises and accounting agencies utilize the internet to carry out intelligent accounting instead of the traditional manual accounting method, and accounting refers to recording all economic services generated by an enterprise and public institution or a personal family on an account book by applying a certain accounting method; the method is characterized in that according to original vouchers and bookkeeping vouchers which are checked to be correct, and accounting subjects specified by a national unified accounting system are used, and a compound bookkeeping method is used for sequentially and classically registering economic businesses in an account book.

However, when enterprises use the internet to carry out intelligent bookkeeping at present, manual input is needed when invoices cannot be automatically identified after being uploaded, manual assistance is needed for bank and cash daily keeping, detailed accounts cannot be automatically generated, tax can not be automatically reported, and bookkeeping vouchers and expense reimbursement notes cannot be automatically generated.

Disclosure of Invention

The invention aims to solve the defects in the prior art and provides an artificial intelligence accounting automatic bookkeeping and tax declaring method and system.

In order to achieve the purpose, the invention is realized by the following technical scheme:

a method and system for automatic accounting and tax return of artificial intelligent accounting is used for realizing the method and system for automatic accounting and tax return in intelligent mobile phones, portable computers and computer equipment through interaction with a big data center and a neural network through a wireless network or 4G and 5G networks, the method and system are characterized in that invoices, bank running water and receipt are shot through a mobile phone camera or automatically acquired through a third-party electronic invoice platform SDK and then transmitted to an application layer through the wireless network or the 4G and 5G networks, intelligent accounting services such as accounting vouchers, detailed accounts and expense reimbursement bills are automatically generated for users based on small enterprise accounting criteria, enterprise accounting criteria and tax laws serving as basic algorithm logic, and tax returns tax automatically; the big data center system conducts training iteration according to application layer data, habit memory data of the user and learning analysis data of the big data center database in the database module through deep learning of a large amount of accounting bookkeeping behaviors, and data of the training iteration are output to the user application layer to help the user to achieve zero-basis financial bookkeeping. The application layer comprises a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax payment module, an accounting firm audit acquisition module and a national tax administration supervision module; the system is provided with a neural network module, a database module and a big data center system module independently.

The system comprises a user registration module, a database module and a neural network module, wherein the user registration module is connected with a sky eye checking SDK (security data reader) after submitting a company name and verifying the company name by a mobile phone verification code, automatically acquires registration capital, an establishment date, unified social credit code, a taxpayer identification number, a company type, a registration authority, industry and operation range information, interacts with the database module and the neural network module based on the basic algorithm logic of small enterprise accounting criteria, enterprise accounting criteria and tax law, memorizes and stores the data and intelligently matches the user industry subject module, tax rate and accounting criteria, and automatically activates a user account and generates an initial account cover; the user registration module is connected with a comprehensive business handling platform SDK of a torch center of the science and technology department, automatically inquires whether a user enterprise is a national high and new technology enterprise or an evaluated small and medium science and technology enterprise, and judges whether a research and development auxiliary account and a research and development expense adding and deduction proportion are independently set for the user according to the acquired data.

The industry subject characteristic module is connected with the database module and the neural network module, is based on the accounting criteria of small enterprises and the accounting criteria of enterprises as basic algorithm logic, is used for memorizing and storing data in user industry subject information, and utilizes the database module and the neural network module to train and learn to realize intelligent matching. The bill feature module is connected with the database module, the neural network module, the industry subject feature module, the electronic certificate feature module, the electronic account book feature module, the electronic invoice module and the research and development auxiliary account module, is based on the accounting criteria of small enterprises, the accounting criteria of enterprises, tax laws and research and development plus deduction policies as basic algorithm logic, and is used for enabling paper invoices uploaded by users, electronic invoices obtained through SDK of third-party electronic invoices and central non-tax income unified bill mainstream bills to identify invoicing dates, tax-containing prices, tax-free prices, invoice numbers, goods or taxable labor, service names, project names, bill issuing parties and bill receiving party information in bills, memorizing and storing data of the bill feature module in the database module and the neural network module, and training and learning are carried out by the database module and the neural network module to realize subject collection, account book feature module, electronic invoice bill, Automatically generating an electronic expense reimbursement bill, automatically generating an electronic accounting voucher and automatically recording data into an electronic account book; the user can carry out manual modification, and the database module memorizes the user accounting habits.

The bank receipt and statement feature module is connected with the database module, the neural network module, the electronic certificate feature module and the electronic account book feature module, a user can export statement of public accounts, receipt uploading or paper printing statement and receipt uploading through an online bank, and the database module and the neural network module automatically generate bank accounts, cash diary accounts and receipt automatic judgment and identification to enter in the corresponding accounting certificates; the user can make manual modification and intervention, and the database module memorizes the user habit.

The electronic certificate feature module is connected with the database module, the neural network module, the electronic account book feature module, the bill feature module, the bank receipt and the statement feature module, is based on the accounting criterion of small enterprises, the accounting criterion of enterprises and tax law as basic algorithm logic, and is used for automatically generating a charge reimbursement note and a certificate for bills and bank receipts uploaded by a user; the user can make manual modification and intervention, and the database module memorizes the user habit. The electronic account book feature module is connected with the database module, the neural network module, the bill feature module, the bank receipt and statement feature module, the electronic certificate feature module, the industry subject feature module, the electronic invoice module and the electronic report module, the neural network module and the database module automatically match the electronic account book for a user on the basis of basic algorithm logic and memory data of small enterprise accounting criteria, tax laws, research and development addition deduction policies, and if the user is a national high and new technology enterprise or a scientific and technological type medium and small enterprise, the automatic user generates a research and development auxiliary electronic account book by matching; the user can make manual modification and intervention, and the database module memorizes the user habit.

And the electronic report module is connected with the electronic account book characteristic module and the database module and is used for automatically generating electronic reports according to the month, the season and the year of the data in the electronic account book of the user.

The financial data analysis module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the bank receipt and statement characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the inventory module, the salary module, the electronic invoice module, the research and development auxiliary account module, the tax payment module, the accounting firm audit access module, the national tax administration supervision module, the database module and the neural network module, and is used for analyzing the financial data of the current month, season and year of the user and giving early warning to the operation risk and the tax risk of the user based on the data memorized and deeply learned and trained by the neural network module, the database module and the large data center system module; the user can look up the electronic accounting voucher, the electronic account book, the research and development auxiliary electronic account book and the electronic report in the current month, the current season and the current year.

The inventory module is connected with the electronic account book characteristic module, the electronic report module, the electronic invoice module, the electronic certificate characteristic module, the database module and the neural network module and is used for stocking, selling and storing commodities stored by a user, and the module is connected with a local smart phone and a portable computer of the user and uses an equipment camera to identify a commodity bar code to make warehousing registration; automatically processing the flows of inventory commodity account, generating purchase and sale receipts, generating sale receipts (requiring a user to install a receipt machine locally), issuing electronic sale receipts and the like for the user; the user can make manual modification and intervention, and the database module memorizes the user habit.

The salary module is connected with the industrial subject characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the financial data analysis module, the inventory module, the salary module, the research and development auxiliary account module, the database module and the neural network module, a user presets the worker information, the salary standard, the reward standard, the attendance standard, the submission standard, the five-insurance one-money standard and the personal tax deduction setting, the attendance data can be imported by a local attendance machine or automatically acquired by being connected with an enterprise WeChat SDK and a bite SDK, and the staff wage, the submission, the wage table generation, the accounting voucher generation, the automatic collection accounting subjects, the automatic personal tax declaration and the annual personal tax remittance clearing are carried out; the user can make manual modification and intervention, and the database module memorizes the user habit.

The electronic invoice module is connected with the user registration module, the industry and subject feature module, the bill feature module, the electronic certificate feature module, the electronic account book feature module, the electronic report module, the bank receipt and statement feature module, the financial data analysis module, the inventory module, the research and development auxiliary account module, the salary module, the database module, the neural network module and the third-party electronic invoice SDK and used for online invoicing of the user, the user only needs to input the name, the amount and the invoice content of a bill recipient in the electronic invoice module, and the electronic invoice module is connected with the user registration module to automatically fill and input a unified social credit code, an address and telephone information; if the person receiving the ticket is the system user, automatically sending an electronic invoice to the account of the user receiving the ticket and automatically executing accounting work; if the person receiving the ticket is not the system user, the mobile phone number or the mailbox of the person receiving the ticket needs to be input to send the electronic invoice or the invoice is sent by the person making the ticket through self-downloading. The electronic invoice module automatically executes accounting work after invoicing; the user can make manual modification and intervention, and the database module memorizes the user habit.

The tax declaration module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the bank receipt and statement characteristic module, the financial data analysis module, the inventory module, the salary module, the research and development auxiliary account module, the national tax administration supervision module, the database module and the neural network module, a tax law is taken as a basic algorithm, the latest tax policy of a website of the national tax administration is acquired in real time through a spider crawler program, and the latest tax policy is matched with the database module and the neural network module to be memorized and learned and is used for value-added tax, acquired tax, personal tax and remittance settlement declaration of a user enterprise; the user can make manual modification and intervention, and the database module memorizes the user habit.

The research and development auxiliary account module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the tax payment declaration module, the database module and the neural network module, and automatically generates a research and development auxiliary account according to a project name and a budget table preset by a user by taking a research and development addition deduction policy as a basic algorithm; the user can make manual modification and intervention, and the database module memorizes the user habit.

The accounting firm audit access module is connected with the user registration module, the electronic certificate feature module, the electronic account book feature module, the electronic report module, the tax payment declaration module, the research and development auxiliary account module, the database module and the neural network module, when the accounting firm provides audit service for users, the accounting firm can access the electronic certificate, the electronic account book, the electronic report, the tax payment declaration form and the research and development auxiliary account by one key, and after the audit work is finished, the accounting firm uploads the electronic audit report; when the accounting affair place is quoted, the database module memorizes the auditing habit of the accounting affair place.

The national tax administration supervision module is connected with a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax declaration module, a neural network module, a database module and a large data center system module, and branch office special managers and tax inspects of the national tax administration can perform real-time online check on electronic invoices, electronic certificates, electronic account books, electronic statements, tax declaration forms and research and development auxiliary accounts of the dominated enterprises or tax risk enterprises pushed by the neural network module, the database module and the large data center system module and issue electronic correction notices to enterprises with problems; the database module memorizes the whole process in real time and is used for learning and training of the neural network module.

The database module is connected with a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement checking characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax declaration module, an accounting firm audit and access module, a national tax administration supervision module and a neural network module, and is provided with a tax database, a big data center database and a neural network deep learning and training database based on the development design of a Mysql database, and the data storage and storage module is used for user registration data, bills, bank receipt and statement checking, electronic certificates, an electronic account book, an electronic report, inventory, salary, electronic invoices, research and development auxiliary accounts, tax declaration, audit and access and national tax administration supervision, Memorizing; learning and training data are memorized through the database module and learned through the neural network module, and an intelligent accounting method which is memorized, learned and trained is output to the application layer for a user.

The neural network module is connected with the database module, is provided with an input neuron X, a hidden layer H and an output neuron Y, and is mainly used for performing training iteration according to application layer data and user habit memory data of a user in the database module, and outputting the data of the training iteration to the user application layer.

The big data center system module is connected with the database module and the neural network module and is mainly used for collecting, transmitting, modeling and storing data, counting, analyzing, learning and mining the data, visualizing and feeding back the data according to the finance and tax data of the database module;

preferably, the tax declaration module uses a JAVA-based spring boot frame to realize a cloud login server, the cloud login server is connected with a gold tax disk and a tax KKEY soft certificate centralized management device, tax declaration time is preset for a user, tax copying work is automatically realized for the user, a region ID, a user name and a password matched with an enterprise login tax office are sent to the tax office server through an http request, and corresponding data packets are automatically sent to the tax office server to finish enterprise value-added tax, enterprise income tax, remittance clearing and individual tax declaration work according to an electronic account book, a report, a wage table and a research and development auxiliary account of the user, so that automatic tax declaration operation is realized.

Preferentially, the neural network module is developed based on Python language, an input neuron X, a hidden layer H and an output neuron Y are developed, deep learning training iteration is carried out according to application layer data and user habit memory data of a user in the database module, and an artificial intelligence accounting method and data subjected to deep learning training are output for the user.

Preferably, the calculation method of the artificial intelligence accounting automatic accounting and tax declaring method and the system thereof comprises the following steps:

s1: the user enters a user registration unit, inputs an enterprise name, automatically checks the SDK with the sky eye to obtain and automatically obtain registration capital, a formation date, a unified social credit code, a taxpayer identification number, a company type, a registration authority, an industry and an operation range information, inputs a preset login password, face information and a mobile phone number, and successfully registers after inputting a short message verification code; the database module and the neural network module memorize, store and intelligently match the user industry subject module, the tax rate and the accounting criteria, and automatically activate a user account and generate an initial account set; and the system is connected with the SDK of the comprehensive business handling platform of the torch center of the science and technology department, automatically inquires whether a user enterprise is a national high and new technology enterprise or an evaluated small and medium science and technology enterprise, and judges whether a research and development auxiliary account and a research and development expense adding and deducting proportion are independently set for the user according to the acquired data.

S2: after the user is successfully registered in step S1, the user enters a user login unit, and then inputs a password or performs face recognition login in a password input module, and a password confirmation module confirms the password or compares the face with the password;

s3: after the password is confirmed to be correct in step S2, the user enters a user management unit, and performs tax payment user name and password presetting, centralized trusting of a gold tax disk and a tax UKEY soft certificate, electronic invoice account setting, enterprise wechat account setting authorization, sting account setting authorization, subject modification, tax rate modification, tax copying and tax return time setting, staff information and salary setting, social security setting, and account cover enabling date setting (if the establishment date is 15 days earlier than the registration date, it is required to import early-stage financial data or make up account registration operation);

s4: after the basic information is set in step S3, the user can check the electronic accounting document, the electronic report, the electronic account book, the tax payment information, the inventory, the manual intervention modification, the finance and tax risk disclosure information, the data backup and download, the printing of the certificate account book report and the one-key audit in real time;

s5: the method comprises the steps that a user clicks a received bill to enter, a paper invoice can be uploaded through photographing, an electronic invoice obtained through a third-party electronic invoice SDK and a central non-tax income unified bill paper or electronic invoice can be uploaded, a system OCR component automatically identifies the invoicing date, the tax value and the tax free, the invoice number, goods or tax-related labor, the service name, the project name, the invoice issuing party and the information of the received bill, after the uploading is finished, the user only needs to select cash payment or bank payment (the bank payment needs to be uploaded to bank payment certificates), and after the operation is finished, the system automatically performs subject collection for the bill uploaded by the user, automatically generates an electronic expense reimbursement bill, automatically generates an electronic accounting certificate and automatically records data into an electronic account book; the user can make manual modification;

s6: the user clicks on the operation interface to make an invoice, the user needs to manually input the company name, the amount, the content, the mobile phone number or the mailbox of the invoice taker, and the system automatically supplements the tax number and the address of the invoice taker for the user; after the invoice is successfully invoiced, the system automatically sends the invoice to an email or a mobile phone of the invoice receiver; the receipt system executes automatic accounting operation; if the ticket receiver is the system user, the user side of the ticket receiver automatically executes accounting operation;

s7, the user can export the statement of public account, receipt, and the receipt uploaded or print the paper statement and receipt through online banking on the operation interface, the database module and the neural network module automatically generate the bank account book, the cash diary account book and the receipt, and automatically judge, identify and enter the corresponding accounting voucher;

s8: the user inventory commodity purchase, sales and inventory management system is connected with a user local smart phone and a portable computer and uses an equipment camera to identify a commodity bar code to make warehousing registration; automatically processing inventory commodity accounts, generating purchase and sale receipts, generating sales tickets (requiring a user to install a ticket machine locally) and issuing electronic sales invoices for the user;

s9: the user salary function can be set according to the user preset staff information, salary standards, punishment standards, attendance standards, promotion standards, five-risk one-money standards and individual tax deduction, attendance data can be imported through a local attendance machine or automatically acquired through connection with an enterprise WeChat SDK and a sting SDK, and wage, promotion, wage table generation, accounting voucher generation, account accounting subjects automatic collection, personal tax declaration and annual individual tax settlement are automatically checked and paid;

s10: if the user is a national science and technology type medium and small enterprise, a national high and new technology enterprise or a project execution unit, the system automatically generates a research and development auxiliary account according to a project name and an expenditure budget table preset by the user;

s11: after the operations of S1, S2, S3, S4, S5, S6, S7, S8, S9 and S10 are completed, the tax payment reporting system is connected with a gold tax disk and a tax UKEY soft certificate centralized management device, tax payment time is preset for a user, tax copying work is automatically realized for the user, a regional ID, a user name and a password matched with an enterprise login tax office are sent to a tax office server through an http request, and corresponding data packets are automatically sent to the tax office server according to an electronic book, a report, a payroll and a research and development auxiliary account of the user to complete enterprise tax value increment, enterprise income tax payment, settlement and individual tax reporting work, so that automatic tax payment operation is realized;

s12: the user can entrust the accounting firm on line in the system to provide auditing service for the user, and the entrusted accounting firm realizes one-key data taking and one-key acquisition of electronic certificates, electronic account books, electronic statements, tax payment statements and research and development auxiliary accounts; according to the online quotation of the user data accounting firm, the user pays corresponding fees to the system platform to host according to the quotation; after the audit work is finished, the accounting firm uploads an electronic audit report, and the system automatically settles with the accounting firm after the user confirms that the audit report is received;

s13: the national supervision end can set account numbers and authorities according to the special administrators of local offices and tax inspection grades of the national tax administration, can perform real-time online spot inspection on electronic invoices, electronic certificates, electronic account books, electronic reports, tax payment reports and research and development auxiliary accounts of the administered enterprises or big data centers and tax risk enterprises pushed by a neural network, and issue electronic correction or administrative penalty notifications to problematic enterprises;

s14: the database is divided into a finance and tax database, a big data center database and a neural network deep learning training database, wherein the finance and tax database is used as a memory storage library and is used for storing and memorizing data of user registration, bills, bank receipt and statement, electronic certificates, electronic account books, electronic statements, inventory, salaries, electronic invoices, research and development assisted accounts, tax declaration, audit data taking and supervision of the national tax administration; the big data center database is used for data modeling, statistics, analysis and output to the neural network deep learning training database; learning training data by a neural network module, and outputting an intelligent accounting method subjected to memory, learning and training to an application layer for a user;

step S15: the neural network training data are obtained from deep learning data summarized according to a large data center database in the database module to perform deep learning training and iteration, and the data of training iteration are output to a user application layer; the method comprises the following specific steps:

step S15-1: extracting data: extracting invoice information, voucher information, subject information and account book information from a big data center database according to user groups;

step S15-2: and (3) training data assembly: setting subjects, certificates and accounts which are matched with invoice information, certificate information and account book information in single user data in a big data center database as correct answers, taking W-1 wrong answers from the remaining subjects, certificates and accounts books, randomly taking 1 to W/2 same-father subjects, certificates and account book options from the wrong answers, randomly taking 1 to W/2 same-root subjects, certificates and account book options from the wrong answers, randomly selecting the rest, and randomly arranging all the options to obtain assembled data;

step S15-3: batch training: dividing all the assembled data according to a set proportion, such as the proportion of 80% training, 10% verification and 10% testing, or the proportion of 85% training, 10% verification and 5% testing; let the model predict the data for a set number of batches, e.g. 300, and cross entropy for the prediction results: (

) Calculating the cross-entropy loss while performing a random gradient descent of the cross-entropy loss for each batch (

) And (4) performing back propagation, wherein each training turn can train 1000 batches, and after each batch of training is finished, the probability and cross entropy loss values are performed by using verification data, and the accuracy is measured.

Step S15-4: and (3) storing training data: and (5) continuously repeating the training turns, observing the verification accuracy and the change of the cross entropy loss value, and storing the optimal training result of the verification performance as a final result.

Step S15-5: training data use: and the finance and tax system acquires the optimal training result and can remind and correct the user when the user inputs errors.

S16: the big data center system module is used for acquiring, transmitting, modeling and storing data, counting, analyzing and mining the data, visualizing and feeding back the data according to the finance and tax data acquired by the database module in real time and in batches;

s17: and the user financial and tax data, the notice of successful tax payment and tax return, the risk early warning notice and the tax administration notice are sent to a user short message, a WeChat applet, an APP terminal and a user management terminal in real time.

Advantageous effects

The invention provides a method and a system for automatic accounting and tax declaration of artificial intelligence accounting, which have the following beneficial effects compared with the prior art:

(1) the system and the method for automatically bookkeeping and tax return in artificial intelligent accounting automatically transmit the obtained invoice, bank running water and receipt through a camera of a user intelligent mobile phone, a camera of a portable computer and a computer device or automatically acquired through a third-party electronic invoice platform SDK to an application layer through a wireless network or a 4G or 5G network, automatically generate intelligent bookkeeping businesses such as bookkeeping vouchers, itemized accounts and expense reimbursement bills for users based on accounting criteria of small enterprises, accounting criteria of enterprises and tax laws as basic algorithm logic, and automatically apply tax declaration in tax period; the neural network helps a user to realize zero-basis financial bookkeeping work through deep learning training of a large amount of accounting bookkeeping behaviors.

(2) The registration module is connected with the sky eye checking SDK to automatically acquire registration capital, establishment date, unified social credit code, taxpayer identification number, company type, registration authority, industry and operation range information, the data interacts with the database module and the neural network module based on the basic algorithm logic of the small enterprise accounting criteria, enterprise accounting criteria and tax law, the database module and the neural network module memorize and intelligently match the user industry subject module, tax rate and accounting criteria, automatically activate the user account and generate an initial account cover; the user registration module is connected with a comprehensive business handling platform SDK of a torch center of the science and technology department, automatically inquires whether a user enterprise is a national high and new technology enterprise or an evaluated small and medium science and technology enterprise, and judges whether a research and development auxiliary account and a research and development expense adding and deduction proportion are independently set for the user according to the acquired data; the user does not need complicated operation, the user experience can be improved, and the simple intelligent finance and tax matching setting method without professional knowledge is provided for the user.

(3) The subject characteristic module is connected with the database module and the neural network module, the data in the subject information of the user industry can be memorized and stored based on the accounting criteria of small enterprises and the accounting criteria of enterprises as basic algorithm logic, the data can be interacted with the database module and the neural network module, and the neural network module is deeply learned, trained and output to an application layer to realize intelligent matching of the subjects of the user.

(4) The manual intelligent accounting automatic bookkeeping and tax return system and the calculation method thereof, a bill characteristic module is used for entering a paper invoice uploaded by a user, an electronic invoice obtained through a third-party electronic invoice SDK and a central non-tax income unified bill mainstream bill, identifying the invoice date, the tax price and the tax free, the invoice number, goods or taxation labor, the service name, the project name, the invoice issuing party and the information of the invoice receiving party in the bill by an OCR component, memorizing and storing the data in a database module and a neural network module by the bill characteristic module, and the database module and the neural network module are used for training and learning to realize the subject collection of the bill, automatically generate an electronic expense return bill, automatically generate an electronic bookkeeping certificate and automatically record data into an electronic account book; helping the user to implement zero-base financial accounting work.

(5) The user can export the statement of public account, receipt, statement uploading or printing paper statement and receipt through the online bank, and the database module and the neural network module automatically generate a bank account book, a cash diary account book and a receipt automatic judgment and identification to be returned to the corresponding accounting voucher; alleviate traditional financial staff's manual work load of typeeing of data, effectively promote financial accounting work efficiency.

(6) The system and the calculation method thereof for the artificial intelligent accounting automatic bookkeeping and tax declaring are characterized in that an inventory module is connected with a user local smart phone and a portable computer and uses an equipment camera to identify a commodity bar code to make warehousing registration; automatically processing the flows of inventory commodity account, generating purchase and sale receipts, generating sale receipts (requiring a user to install a receipt machine locally), issuing electronic sale receipts and the like for the user; the problem that traditional inventory management personnel need use the complicated operation of manual type-in and bar code gun type-in is solved, the working efficiency of the inventory management personnel is improved, and the management flow is simplified.

(7) The salary module is set as a basic algorithm according to the worker information, the salary standard, the punishment standard, the attendance standard, the promotion standard, the five-risk one-money standard and the personal tax deduction preset by a user, the attendance data is imported through a local attendance machine or is automatically acquired through being connected with an enterprise WeChat SDK and a sting SDK, and the wage, the calculation, the wage table generation, the bookkeeping voucher generation, the automatic collection and accounting subjects, the automatic personal tax declaration and the annual personal tax remittance and payment are automatically checked; effectively improve financial staff wage accounting efficiency.

(8) The electronic invoice module is connected with a third-party electronic invoice SDK and used for online invoicing of a user, the user only needs to input the name, the amount and the invoice content of an enterprise of a receiver in the electronic invoice module, and the electronic invoice module is connected with a user registration module to automatically fill and input unified social credit code, address and telephone information; if the person receiving the ticket is the system user, automatically sending an electronic invoice to the account of the user receiving the ticket and automatically executing accounting work; if the person receiving the ticket is not the system user, the mobile phone number or the mailbox of the person receiving the ticket needs to be input to send the electronic invoice or the invoice is sent by the person making the ticket through self-downloading. The electronic invoice module automatically executes accounting work after invoicing; the method effectively forms an electronic invoice using closed loop, realizes paperless financial tax and reduces the workload of financial staff.

(9) The artificial intelligent accounting automatic bookkeeping and tax return system and the calculating method thereof are characterized in that a tax return reporting module uses a JAVA-based spring boot frame to realize a cloud login server, the cloud login server is connected with a gold tax disk and a tax KKEY soft certificate centralized management device, tax return work is automatically realized for a user at the preset tax return time of the user, the matched area ID, user name and password of an enterprise login tax office are sent to a tax office server through an http request, and corresponding data packets are automatically sent to the tax office server to finish the enterprise value-added tax, enterprise income tax, settlement and payment and individual tax return work according to an electronic account book, a report, a wage table and research and development auxiliary account of the user, so that the automatic tax return reporting operation is realized. The problem that the traditional manual work is easy to miss and mistaken reports is solved.

(10) The neural network carries out deep learning training iteration according to a finance and tax database, a big data center database and a neural network deep learning training database of a user in the database module, and data of the training iteration is output to a user application layer; the method solves the problem of subject collection, tax declaration, electronic account book and accounting voucher in accounting of users of different industry types, provides a simple and convenient method for automatically accounting and declaring tax for the users, improves the financial and tax efficiency of the users and does not need the users to master professional financial and tax knowledge.

Drawings

FIG. 1 is a schematic frame diagram of the present invention;

FIG. 2 is a schematic block diagram of a neuron according to the present invention;

FIG. 3 is a schematic block diagram of an all neural network of the present invention;

FIG. 4 is a schematic frame diagram of neural network invoice information deep learning training of the present invention;

FIG. 5 is a schematic framework diagram of neural network billing voucher deep learning training of the present invention;

FIG. 6 is a schematic frame diagram of neural network account book information deep learning training of the present invention;

FIG. 7 is a schematic frame diagram of the neural network deep learning training of the present invention;

FIG. 8 is a schematic frame diagram of the three-layer fully-connected neural network deep learning training of the present invention;

FIG. 9 is a block diagram of a schematic framework for user registration in accordance with the present invention;

FIG. 10 is a schematic block diagram of a user base arrangement of the present invention;

FIG. 11 is a schematic frame diagram of the industrial subject features of the present invention;

FIG. 12 is a schematic frame diagram of a note feature of the present invention;

FIG. 13 is a conceptual framework diagram of the bank receipt and statement feature of the present invention;

FIG. 14 is a conceptual framework diagram of the electronic voucher feature of the present invention;

FIG. 15 is a conceptual framework diagram of the electronic book feature of the present invention;

FIG. 16 is a conceptual framework diagram of an electronic report of the present invention;

FIG. 17 is a conceptual framework diagram of financial data analysis according to the present invention;

FIG. 18 is a schematic frame diagram of the inventory of the present invention;

FIG. 19 is a schematic block diagram of a compensation module according to the present invention;

FIG. 20 is a schematic frame diagram of an electronic invoice according to the present invention;

FIG. 21 is a conceptual framework diagram of a development aid book of the present invention;

FIG. 22 is a conceptual framework diagram of tax declaration according to the present invention;

FIG. 23 is a schematic block diagram of an audit access by an accounting firm according to the present invention;

FIG. 24 is a schematic block diagram of the national tax administration supervision of the present invention;

FIG. 25 is a schematic frame diagram of the database module of the present invention;

FIG. 26 is a schematic block diagram of a big data center system of the present invention;

Detailed Description

To further clarify the objects, structures, features and functions of the present invention, preferred embodiments are described in detail below.

As shown in fig. 1, an artificial intelligence accounting automatic bookkeeping and tax return system is used for interacting with a big data center and a neural network at a user application layer and realizing intelligent bookkeeping services such as accounting automatic bookkeeping, tax return, generation of bookkeeping vouchers, itemized accounts, cost reimbursement notes and the like, and tax return is automatically executed in a tax return month; the big data center system conducts training iteration according to application layer data, habit memory data of the user and learning analysis data of the big data center database in the database module through deep learning of a large amount of accounting bookkeeping behaviors, and data of the training iteration are output to the user application layer to help the user to achieve zero-basis financial bookkeeping. The application layer comprises a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax payment module, an accounting firm audit acquisition module and a national tax administration supervision module; the system is provided with a neural network module, a database module and a big data center system module independently.

As shown in fig. 2-8, the neural network module includes an input neuron X, a hidden layer H, an output neuron Y, a neural network subject training module, a billing voucher training module, and an account book training module; the neural network module is connected with the database module and the application layer module and used for extracting invoice information, certificate information, subject information and account book information from a large data center database in the database module, training and assembling the extracted data, training the assembled data according to a set proportion and batch number, performing probability and cross entropy loss value and measuring accuracy by using verification data after each batch of training is finished, continuously repeating training rounds, observing the verification accuracy and the change of the cross entropy loss value, and storing the optimal training result of verification performance as a final result; and the application layer acquires the optimal training result and helps the user to realize zero-base financial accounting.

As shown in fig. 9, the user registration module includes a user registration unit and an SDK, is connected to the application layer module and the database module, and is used for password presetting, face entry, initial information acquisition, and account set matching functions during user registration, thereby simplifying the flow of user registration and setting.

As shown in fig. 10, the user basic settings include user login, face and password verification comparison, SDK, user management unit, which is connected to the application layer, the fiscal database, the tax UKEY soft certificate centralized management device, and the gold tax disk centralized management device, and is used for user tax payment user name and password presetting, gold tax disk and tax UKEY soft certificate centralized hosting entrustment, electronic invoice account setting, ticketing and billing functions, inventory management, third party attendance setting, research and development auxiliary account setting, subject management, tax rate management, tax copying and tax reporting time setting, employee information and pay setting, social security setting, account set enabling date setting, financial audit, data backup and download, and voucher book report printing; the user can check and manually intervene the electronic accounting document, the electronic report form, the electronic account book, the tax payment information, the inventory and the tax risk disclosure, and the intervention result and the reason are transmitted to the neural network for deep learning training.

As shown in fig. 11, the industry subject feature module is connected to the database module, and is configured to obtain industry subject features stored in the finance and tax database, and the application layer is configured to automatically determine user industry features and match industry subjects when performing accounting operation.

As shown in FIG. 12, the ticket feature module includes

As shown in FIG. 13, the bank receipt and statement feature module includes

As shown in FIG. 14, the electronic voucher feature module comprises

As shown in FIG. 15, the electronic book features module includes

As shown in FIG. 16, the electronic report features module includes

As shown in FIG. 17, the financial data analysis module includes

As shown in FIG. 18, the inventory module includes

As shown in FIG. 19, the compensation module includes

As shown in FIG. 20, the electronic invoice module includes

As shown in FIG. 21, the development support ledger module comprises,

As shown in FIG. 22, the tax declaration module includes

As shown in FIG. 23, the accounting firm audit module comprises

As shown in FIG. 24, the State tax administration module includes

As shown in FIG. 25, the database module includes

As shown in FIG. 26, the big data center system module includes

Further, the calculation method of the artificial intelligent accounting automatic bookkeeping and tax declaring system comprises the following steps:

step S1: the user enters a user registration unit, inputs an enterprise name, automatically checks the SDK with the sky eye to obtain and automatically obtain registration capital, a formation date, a unified social credit code, a taxpayer identification number, a company type, a registration authority, an industry and an operation range information, inputs a preset login password, face information and a mobile phone number, and successfully registers after inputting a short message verification code; the database module and the neural network module memorize, store and intelligently match the user industry subject module, the tax rate and the accounting criteria, and automatically activate a user account and generate an initial account set; and the system is connected with the SDK of the comprehensive business handling platform of the torch center of the science and technology department, automatically inquires whether a user enterprise is a national high and new technology enterprise or an evaluated small and medium science and technology enterprise, and judges whether a research and development auxiliary account and a research and development expense adding and deducting proportion are independently set for the user according to the acquired data.

Step S2: after the user is successfully registered in step S1, the user enters a user login unit, and then inputs a password or performs face recognition login in a password input module, and a password confirmation module confirms the password or compares the face with the password;

step S3: after the password is confirmed to be correct in step S2, the user enters a user management unit, and performs tax payment user name and password presetting, centralized trusting of a gold tax disk and a tax UKEY soft certificate, electronic invoice account setting, enterprise wechat account setting authorization, sting account setting authorization, subject modification, tax rate modification, tax copying and tax return time setting, staff information and salary setting, social security setting, and account cover enabling date setting (if the establishment date is 15 days earlier than the registration date, it is required to import early-stage financial data or make up account registration operation);

step S4: after the basic information is set in step S3, the user can check the electronic accounting document, the electronic report, the electronic account book, the tax payment information, the inventory, the manual intervention modification, the finance and tax risk disclosure information, the data backup and download, the printing of the certificate account book report and the one-key audit in real time;

step S5: the method comprises the steps that a user clicks a received bill to enter, a paper invoice can be uploaded through photographing, an electronic invoice obtained through a third-party electronic invoice SDK and a central non-tax income unified bill paper or electronic invoice can be uploaded, a system OCR component automatically identifies the invoicing date, the tax value and the tax free, the invoice number, goods or tax-related labor, the service name, the project name, the invoice issuing party and the information of the received bill, after the uploading is finished, the user only needs to select cash payment or bank payment (the bank payment needs to be uploaded to bank payment certificates), and after the operation is finished, the system automatically performs subject collection for the bill uploaded by the user, automatically generates an electronic expense reimbursement bill, automatically generates an electronic accounting certificate and automatically records data into an electronic account book; the user can make manual modification;

step S6: the user clicks on the operation interface to make an invoice, the user needs to manually input the company name, the amount, the content, the mobile phone number or the mailbox of the invoice taker, and the system automatically supplements the tax number and the address of the invoice taker for the user; after the invoice is successfully invoiced, the system automatically sends the invoice to an email or a mobile phone of the invoice receiver; the receipt system executes automatic accounting operation; if the ticket receiver is the system user, the user side of the ticket receiver automatically executes accounting operation;

step S7, the user can export the statement of public account, receipt, the statement upload or the paper statement and receipt by the online bank on the operation interface, the database module and the neural network module automatically generate the bank account book, the cash diary account book and the receipt, and automatically judge, identify and enter the corresponding accounting voucher;

step S8: the user inventory commodity purchase, sales and inventory management system is connected with a user local smart phone and a portable computer and uses an equipment camera to identify a commodity bar code to make warehousing registration; automatically processing inventory commodity accounts, generating purchase and sale receipts, generating sales tickets (requiring a user to install a ticket machine locally) and issuing electronic sales invoices for the user;

step S9: the user salary function can be set according to the user preset staff information, salary standards, punishment standards, attendance standards, promotion standards, five-risk one-money standards and individual tax deduction, attendance data can be imported through a local attendance machine or automatically acquired through connection with an enterprise WeChat SDK and a sting SDK, and wage, promotion, wage table generation, accounting voucher generation, account accounting subjects automatic collection, personal tax declaration and annual individual tax settlement are automatically checked and paid;

step S10: if the user is a national science and technology type medium and small enterprise, a national high and new technology enterprise or a project execution unit, the system automatically generates a research and development auxiliary account according to a project name and an expenditure budget table preset by the user;

step S11: after the operations of S1, S2, S3, S4, S5, S6, S7, S8, S9 and S10 are completed, the tax payment reporting system is connected with a gold tax disk and a tax UKEY soft certificate centralized management device, tax payment time is preset for a user, tax copying work is automatically realized for the user, a regional ID, a user name and a password matched with an enterprise login tax office are sent to a tax office server through an http request, and corresponding data packets are automatically sent to the tax office server according to an electronic book, a report, a payroll and a research and development auxiliary account of the user to complete enterprise tax value increment, enterprise income tax payment, settlement and individual tax reporting work, so that automatic tax payment operation is realized;

step S12: the user can entrust the accounting firm on line in the system to provide auditing service for the user, and the entrusted accounting firm realizes one-key data taking and one-key acquisition of electronic certificates, electronic account books, electronic statements, tax payment statements and research and development auxiliary accounts; according to the online quotation of the user data accounting firm, the user pays corresponding fees to the system platform to host according to the quotation; after the audit work is finished, the accounting firm uploads an electronic audit report, and the system automatically settles with the accounting firm after the user confirms that the audit report is received;

step S13: the national supervision end can set account numbers and authorities according to the special administrators of local offices and tax inspection grades of the national tax administration, can perform real-time online spot inspection on electronic invoices, electronic certificates, electronic account books, electronic reports, tax payment reports and research and development auxiliary accounts of the administered enterprises or big data centers and tax risk enterprises pushed by a neural network, and issue electronic correction or administrative penalty notifications to problematic enterprises;

step S14: the database is divided into a finance and tax database, a big data center database and a neural network deep learning training database, wherein the finance and tax database is used as a memory storage library and is used for storing and memorizing data of user registration, bills, bank receipt and statement, electronic certificates, electronic account books, electronic statements, inventory, salaries, electronic invoices, research and development assisted accounts, tax declaration, audit data taking and supervision of the national tax administration; the big data center database is used for data modeling, statistics, analysis and output to the neural network deep learning training database; learning training data by the neural network module, and outputting an intelligent accounting method subjected to memory, learning and training to an application layer for a user.

Step S16: the big data center system module is used for acquiring, transmitting, modeling and storing data, counting, analyzing, learning and mining the data, and visualizing and feeding back the data according to the finance and tax data acquired by the database module;

step S17: and the user financial and tax data, the notice of successful tax payment and tax return, the risk early warning notice and the tax administration notice are sent to a user short message, a WeChat applet, an APP terminal and a user management terminal in real time.

And those not described in detail in this specification are well within the skill of those in the art. It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims

1. An artificial intelligent accounting automatic bookkeeping and tax return system is used for interacting with a big data center and a neural network through a wireless network or a 4G and 5G network in a smart phone, a portable computer and computer equipment and realizing the method and the system for accounting automatic bookkeeping and tax return, an invoice, a bank running water and a receipt are shot through a mobile phone camera or are automatically acquired through a third-party electronic invoice platform SDK and then are transmitted to an application layer through the wireless network or the 4G and 5G network, intelligent bookkeeping services such as a bookkeeping voucher, an itemized account and a charge reimbursement note are automatically generated for a user based on the small enterprise accounting criterion, the enterprise accounting criterion and a tax law as basic algorithm logic, and tax return is automatically executed after tax return; the big data center system conducts training iteration according to application layer data, habit memory data of the user and learning analysis data of the big data center database in the database module through deep learning of a large amount of accounting bookkeeping behaviors, and data of the training iteration are output to the user application layer to help the user to achieve zero-basis financial bookkeeping. The method is characterized in that: the system comprises a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax payment module, an accounting firm audit acquisition module and a national tax administration supervision module; the system is provided with a neural network module, a database module and a big data center system module independently.

2. The system for artificial intelligence accounting automatic billing and tax return as claimed in claim 1, wherein: the system comprises a user registration module, a database module and a neural network module, wherein the user registration module is connected with a sky eye checking SDK (security data reader) after submitting a company name and verifying the company name by a mobile phone verification code, automatically acquires registration capital, an establishment date, unified social credit code, a taxpayer identification number, a company type, a registration authority, industry and operation range information, interacts with the database module and the neural network module based on the basic algorithm logic of small enterprise accounting criteria, enterprise accounting criteria and tax law, memorizes and stores the data and intelligently matches the user industry subject module, tax rate and accounting criteria, and automatically activates a user account and generates an initial account cover; the user registration module is connected with a comprehensive business handling platform SDK of a torch center of the science and technology department, automatically inquires whether a user enterprise is a national high and new technology enterprise or an evaluated small and medium science and technology enterprise, and judges whether a research and development auxiliary account and a research and development expense adding and deduction proportion are independently set for the user according to the acquired data.

3. The system for artificial intelligence accounting automatic billing and tax return as claimed in claim 2, wherein: the industry subject characteristic module is connected with the database module and the neural network module, is based on the accounting criteria of small enterprises and the accounting criteria of enterprises as basic algorithm logic, is used for memorizing and storing data in user industry subject information, and utilizes the database module and the neural network module to train and learn to realize intelligent matching.

4. The system for artificial intelligence accounting automatic billing and tax return as claimed in claim 3, wherein: the bill feature module is connected with the database module, the neural network module, the industry subject feature module, the electronic certificate feature module, the electronic account book feature module, the electronic invoice module and the research and development auxiliary account module, is based on the accounting criteria of small enterprises, the accounting criteria of enterprises, tax laws and research and development plus deduction policies as basic algorithm logic, and is used for enabling paper invoices uploaded by users, electronic invoices obtained through SDK of third-party electronic invoices and central non-tax income unified bill mainstream bills to identify invoicing dates, tax-containing prices, tax-free prices, invoice numbers, goods or taxable labor, service names, project names, bill issuing parties and bill receiving party information in bills, memorizing and storing data of the bill feature module in the database module and the neural network module, and training and learning are carried out by the database module and the neural network module to realize subject collection, account book feature module, electronic invoice bill, Automatically generating an electronic expense reimbursement bill, automatically generating an electronic accounting voucher and automatically recording data into an electronic account book; the user can carry out manual modification, and the database module memorizes the user accounting habits.

5. The system of claim 4, wherein the system comprises: the bank receipt and statement feature module is connected with the database module, the neural network module, the electronic certificate feature module and the electronic account book feature module, a user can export statement of public accounts, receipt uploading or paper printing statement and receipt uploading through an online bank, and the database module and the neural network module automatically generate bank accounts, cash diary accounts and receipt automatic judgment and identification to enter in the corresponding accounting certificates; the user can make manual modification and intervention, and the database module memorizes the user habit.

6. The system for artificial intelligence accounting automatic billing and tax return as claimed in claim 5, wherein: the electronic certificate feature module is connected with the database module, the neural network module, the electronic account book feature module, the bill feature module, the bank receipt and the statement feature module, is based on the accounting criterion of small enterprises, the accounting criterion of enterprises and tax law as basic algorithm logic, and is used for automatically generating a charge reimbursement note and a certificate for bills and bank receipts uploaded by a user; the user can make manual modification and intervention, and the database module memorizes the user habit.

7. The system of claim 6, wherein the system comprises: the electronic account book feature module is connected with the database module, the neural network module, the bill feature module, the bank receipt and statement feature module, the electronic certificate feature module, the industry subject feature module, the electronic invoice module and the electronic report module, the neural network module and the database module automatically match the electronic account book for a user on the basis of basic algorithm logic and memory data of small enterprise accounting criteria, tax laws, research and development addition deduction policies, and if the user is a national high and new technology enterprise or a scientific and technological type medium and small enterprise, the automatic user generates a research and development auxiliary electronic account book by matching; the user can make manual modification and intervention, and the database module memorizes the user habit.

8. The system of claim 7, wherein the system comprises: and the electronic report module is connected with the electronic account book characteristic module and the database module and is used for automatically generating electronic reports according to the month, the season and the year of the data in the electronic account book of the user.

9. The system of claim 8, wherein the system comprises: the financial data analysis module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the bank receipt and statement characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the inventory module, the salary module, the electronic invoice module, the research and development auxiliary account module, the tax payment module, the accounting firm audit access module, the national tax administration supervision module, the database module and the neural network module, and is used for analyzing the financial data of the current month, season and year of the user and giving early warning to the operation risk and the tax risk of the user based on the data memorized and deeply learned and trained by the neural network module, the database module and the large data center system module; the user can look up the electronic accounting voucher, the electronic account book, the research and development auxiliary electronic account book and the electronic report in the current month, the current season and the current year.

10. The system of claim 9, wherein the system comprises: the inventory module is connected with the electronic account book characteristic module, the electronic report module, the electronic invoice module, the electronic certificate characteristic module, the database module and the neural network module and is used for stocking, selling and storing commodities stored by a user, and the module is connected with a local smart phone and a portable computer of the user and uses an equipment camera to identify a commodity bar code to make warehousing registration; automatically processing the flows of inventory commodity account, generating purchase and sale receipts, generating sale receipts (requiring a user to install a receipt machine locally), issuing electronic sale receipts and the like for the user; the user can make manual modification and intervention, and the database module memorizes the user habit.

11. The system of claim 10, wherein the system comprises: the salary module is connected with the industrial subject characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the financial data analysis module, the inventory module, the salary module, the research and development auxiliary account module, the database module and the neural network module, a user presets the worker information, the salary standard, the reward standard, the attendance standard, the submission standard, the five-insurance one-money standard and the personal tax deduction setting, the attendance data can be imported by a local attendance machine or automatically acquired by being connected with an enterprise WeChat SDK and a bite SDK, and the staff wage, the submission, the wage table generation, the accounting voucher generation, the automatic collection accounting subjects, the automatic personal tax declaration and the annual personal tax remittance clearing are carried out; the user can make manual modification and intervention, and the database module memorizes the user habit.

12. The system of claim 11, wherein the system comprises: the electronic invoice module is connected with the user registration module, the industry and subject feature module, the bill feature module, the electronic certificate feature module, the electronic account book feature module, the electronic report module, the bank receipt and statement feature module, the financial data analysis module, the inventory module, the research and development auxiliary account module, the salary module, the database module, the neural network module and the third-party electronic invoice SDK and used for online invoicing of the user, the user only needs to input the name, the amount and the invoice content of a bill recipient in the electronic invoice module, and the electronic invoice module is connected with the user registration module to automatically fill and input a unified social credit code, an address and telephone information; if the person receiving the ticket is the system user, automatically sending an electronic invoice to the account of the user receiving the ticket and automatically executing accounting work; if the person receiving the ticket is not the system user, the mobile phone number or the mailbox of the person receiving the ticket needs to be input to send the electronic invoice or the invoice is sent by the person making the ticket through self-downloading. The electronic invoice module automatically executes accounting work after invoicing; the user can make manual modification and intervention, and the database module memorizes the user habit.

13. The system of claim 12, wherein the system comprises: the tax declaration module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the bank receipt and statement characteristic module, the financial data analysis module, the inventory module, the salary module, the research and development auxiliary account module, the national tax administration supervision module, the database module and the neural network module, a tax law is taken as a basic algorithm, the latest tax policy of a website of the national tax administration is acquired in real time through a spider crawler program, and the latest tax policy is matched with the database module and the neural network module to be memorized and learned and is used for value-added tax, acquired tax, personal tax and remittance settlement declaration of a user enterprise; the user can make manual modification and intervention, and the database module memorizes the user habit.

14. The system of claim 13, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the research and development auxiliary account module is connected with the user registration module, the industry subject characteristic module, the bill characteristic module, the electronic certificate characteristic module, the electronic account book characteristic module, the electronic report module, the tax payment declaration module, the database module and the neural network module, and automatically generates a research and development auxiliary account according to a project name and a budget table preset by a user by taking a research and development addition deduction policy as a basic algorithm; the user can make manual modification and intervention, and the database module memorizes the user habit.

15. The system of claim 14, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the accounting firm audit access module is connected with the user registration module, the electronic certificate feature module, the electronic account book feature module, the electronic report module, the tax payment declaration module, the research and development auxiliary account module, the database module and the neural network module, when the accounting firm provides audit service for users, the accounting firm can access the electronic certificate, the electronic account book, the electronic report, the tax payment declaration form and the research and development auxiliary account by one key, and after the audit work is finished, the accounting firm uploads the electronic audit report; when the accounting affair place is quoted, the database module memorizes the auditing habit of the accounting affair place.

16. The system of claim 15, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the national tax administration supervision module is connected with a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax declaration module, a neural network module, a database module and a large data center system module, and branch office special managers and tax inspects of the national tax administration can perform real-time online check on electronic invoices, electronic certificates, electronic account books, electronic statements, tax declaration forms and research and development auxiliary accounts of the dominated enterprises or tax risk enterprises pushed by the neural network module, the database module and the large data center system module and issue electronic correction notices to enterprises with problems; the database module memorizes the whole process in real time and is used for learning and training of the neural network module.

17. The system of claim 16, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the database module is connected with a user registration module, an industry subject characteristic module, a bill characteristic module, a bank receipt and statement checking characteristic module, an electronic certificate characteristic module, an electronic account book characteristic module, an electronic report module, a financial data analysis module, an inventory module, a salary module, an electronic invoice module, a research and development auxiliary account module, a tax declaration module, an accounting firm audit and access module, a national tax administration supervision module and a neural network module, and is provided with a tax database, a big data center database and a neural network deep learning and training database based on the development design of a Mysql database, and the data storage and storage module is used for user registration data, bills, bank receipt and statement checking, electronic certificates, an electronic account book, an electronic report, inventory, salary, electronic invoices, research and development auxiliary accounts, tax declaration, audit and access and national tax administration supervision, Memorizing; learning and training data are memorized through the database module and learned through the neural network module, and an intelligent accounting method which is memorized, learned and trained is output to the application layer for a user.

18. The system of claim 19, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the neural network module is connected with the database module, is provided with an input neuron X, a hidden layer H and an output neuron Y, and is mainly used for performing training iteration according to application layer data and user habit memory data of a user in the database module, and outputting the data of the training iteration to the user application layer.

19. The system of claim 18, wherein the artificial intelligence accounting automatic billing and tax return system comprises: the big data center system module is connected with the database module and the neural network module and is mainly used for collecting, transmitting, modeling and storing data, counting, analyzing, learning and mining the data, visualizing and feeding back the data according to the finance and tax data of the database module in real time and in batches.

20. An artificial intelligence accounting automatic bookkeeping and tax declaring method comprises the following specific steps:

step S15-3: batch training: dividing all the assembled data according to a set proportion, such as the proportion of 80% training, 10% verification and 10% testing, or the proportion of 85% training, 10% verification and 5% testing; having the model predict a set number of data, e.g. 300, per batch and cross entropy the prediction

Calculating cross entropy loss, and simultaneously carrying out random gradient reduction on the cross entropy loss by each batch

And (4) performing back propagation, wherein each training turn can train 1000 batches, and after each batch of training is finished, the probability and cross entropy loss values are performed by using verification data, and the accuracy is measured.