Team Project: AI Coach

Item Function Specifications

Background of the project

Our project is to solve the problem for beginners to use traditional teaching methods have difficulty in learning a programming language, an error can not be timely, easy to understand feedback. Existing programming education system has many drawbacks, such as teaching methods rely on teachers, teaching assistants based, students mistakes strange, can not be resolved, it is difficult to get timely feedback, in turn, long-term repetitive work and also to teachers assistant wasted a lot of time and energy. For students who self-study, relying on books or written materials to learn difficult to achieve the effect of improving the practical ability.

User scene analysis

first name Freshman novice programmer Xiao Ming
programming ability Before going to university played only basic phone, not how access to computers, but often water group chat, typing
motivation This school has required courses, become also important for their professional
purpose Python can learn better solve some basic questions
difficult He did not programmed basis, subject do not move, little grasp of programming concepts, see the error not understand the interpreter
User preferences Hope to have more friendly error, the difficulty of the subject more gradual
The proportion of users Accounting for about 70% of students in the same session
Typical scene Apart from teaching in the lessons and practice program title on the platform, to consolidate the theoretical knowledge learned in the classroom to try to write more efficient and less error-prone code

Product outlook

Computers have become an integral part of the functioning of society drive a technology education program now has been extended to professionals from more ordinary people, it has a lot of users have the desire to learn programming, online media reports quoted: "According to public data online statistics, the first half of 2019, a total of nine domestic children's programming-related projects have been announced the completion of financing, the amount ranging from one million to one hundred million yuan. ", showing that educational programming now get a lot of attention.

Regardless of the method is to rely on teachers, teaching assistants based teaching methods, or self-study, for this type of beginner inexperienced, there is no AI-based learning platform, making the learning process more intelligent, so that the user has made the process of learning to program the error can be accurately corrected.

Therefore, the use of AI to improve the learning experience for beginners to learn programming, as well as the teachers and teaching assistants freed from duplication of effort, the aid program is an excellent idea of ​​education, has a very broad product prospects.

Killer feature

Our killer feature is the personalized recommendation of topics, from less to more topics and real-time modes of friendly error.

By personalizing the subject of recommendation, and the consolidation of user error to compensate in the blind spot of knowledge in programming learning, by the difficulty of control, it can also smooth the learning curve.

By from less to more topics designed to allow users to fill out a small amount of code, the language familiar to the entire frame, in a gradual transition to a block of code completion, completion function, write the entire program, so that the user is doing while also learning problems language features, reduce the difficulties future programming.

Through real-time error-friendly, allowing users to automatically correct grammatical error code as prompted, correct logic may be incorrect code, the code retrieves real-time, but also allows users to understand the syntax features do not come into contact with, from high-quality of learning other people's code, to write more efficient code.

Work Plan

Alpha stage, Model working group is expected to be carried out:

Determining the division (6 hours)

Programming data collection task (10 hours)

Programming data processing tasks (10 hours)

Research topics recommended model (6 hours)

Real-time model code for research tips (6 hours)

Research friendly error message model (6 hours)

Project Topic Recommendation Model (10 hours)

Real-time design code hints for model (10 hours)

Design friendly error message model (15 hours)

Within each set of tests (10 hours)

Interfacing with other groups (10 hours)

Project prototyping

The project is mainly determined prototyping users of the site carried out in what way and when entering the site interactive, and we can guide and prompt the user base in those places.

Prototyping and highlights related to this project is divided into three pages, others, such as login, registration, password recovery and other pages do not skip the detailed design:

 

 

 

 

Guess you like

Origin www.cnblogs.com/jiyan-he/p/11722591.html