We have upgraded the existing live streaming technology to a certain extent. As far as the product form is concerned, we are most similar to Zoom and Skype. But unlike them, our software application is designed to meet the needs of teaching.
Krishna sat in his office at HSR Layout in Bangalore and talked to us about their live streaming technology. The company's conference room design style was inspired by the movie Star Wars and has a bit of a futuristic feel.
In fact, the key to Vedantu's success is simple: They have optimized streaming live streaming technology, which has significantly improved the user experience of online teaching. Now even in areas with poor network signals, students can see smooth live broadcasts.
In addition, in the live broadcast room, many contents can also be presented in real time. For example, the teacher's voice, the teaching video played by the teacher in the classroom, and even the teacher's blackboard, all of which can be displayed on the student's terminal device.
When it comes to blackboard writing, here is a few more words. Strictly speaking, online blackboard writing does not allow students to see the whole process of teachers writing with chalk like offline courses. Its presentation is actually more like that the text of a PPT courseware appears instantaneously as a whole sentence. Vendantu uses a control device to monitor the teacher's blackboard writing. After the instructor finishes writing a complete sentence, the device transmits the written content to the terminal device in the hands of the student. On the student side, the interface of the blackboard is a whiteboard. Whenever the writing on the blackboard is updated, they will see a line of text popping up on the whiteboard.
If the blackboard interface is too large, the teacher will not be visible on the screen. Therefore, in order not to affect the overall learning experience, we have designed the size of the whiteboard to be smaller. Of course, this size is not so small that students cannot see it In fact, from the perspective of students' devices, the whiteboard is just a static pattern. It does not move. What moves are the teacher's gestures, the pictures in the teaching videos, and the increasing content on the whiteboard. Krishna introduced.
Another educational technology company engaged in streaming media livestream optimization business is Unacademy. The difference between them and Vendantu is that the company's focus is not on optimizing video smoothness, but on improving product features.
For teachers who teach online, the biggest contribution of Unacademy, founded in 2015, is that they make content creation easier.
In fact, many people have the desire to share knowledge, but due to the product's functional limitations, for a long time, only a few people can become 'online knowledge sharers.' These people also need to configure their own microphones, The lecture process and PPT content were recorded, and the program was edited repeatedly ... When all these tasks were completed and they needed to upload content, they would find that the most suitable platform is YouTube. However, as we all know, there are too many types of videos on YouTube The creation of new bloggers is difficult to find quickly by target users, so these teachers are actually worried about the distribution of the content. Said Hemesh Singh, founder and CTO of Unacademy.
What Unacademy does is simplify these tedious programs. Teachers can now use the built-in camera on their phones to directly record course content; PPT handouts can not only be easily uploaded to the app, creators can also modify the content of the handout directly in the app, mark key sentences, and even Enter narration. And before sending to students, all audio and video content will be uploaded and stored in the Amazon cloud server, which can also ensure the security of the file content.
A lot of competitive test makers and publishers in the market are now using Unacademy for online teaching because the whole operation process is convenient and fast, Singh added.
At present, more than 10,000 educators have posted content on the Unacademy platform, and the total number of videos has exceeded 1 million. Just recently, the startup announced that they have completed a $ 110 million funding round with investors including Facebook, Transatlantic Investment Group, Sequoia India, Nexus Venture Partners, Steadview Capital, Blume Ventures, and more.
Leverage Edu's LinkedIn account describes his business very clearly:
This platform is like a market. Here, the machine learning mechanism allows every visiting classmate to meet professionals, and experts are like counselors: he helps you with career planning and helps you apply to the right university , He makes you fully prepared before entering the workplace. In a word, he is your mentor and even your friend.
In a noisy cafe on Lodhi Road in Delhi, Akshay Chaturvedi, founder and CEO of Leverage Edu, tells us more about the company.
It is reported that 750,000 people visit the site every month. Most of these visitors want to try the exact matching function of the platform: through AI tools, users who want to apply for foreign universities can find schools that match their various abilities and personality traits. Students who are worried about choosing a major will also find relevant professional courses that are suitable for them and can stimulate their potential.
But isn't this feature too mysterious? How exactly did Leverage do it?
The answer is simple: data. Leverage's intelligent matching algorithm is implemented with the support of massive data.
When doing machine learning training, we entered different types of schools and data that people have selected in the past decade. Chaturvedi said.
In addition, this algorithm also takes into account the background information of scholars. This data includes school experience, graduation exam scores, work experience, and possibly even employer information. Each data indicator has a different weight.
Of course, we also have a lot of data from universities. For example, compared to some applicants who work for large enterprises, many universities today pay more and more attention to those who have had work experience in startups. So our database This trend must also be tracked and echoed. Chaturvedi added, At present, our matching accuracy has reached 80 ~ 85%. Of course, this is still not enough, considering that the algorithm can never reach 100% accuracy, We decided to include artificial factors as well: expert consultants hired by the company can help users through offline consultation, and do our best to help scholars find the school that suits him or her best.
Leverage can also predict future career choices of users through AI.
This is one of our many studios. In this project, the algorithm is mainly driven by relevant data from clinical psychology. It analyzes your career choices by testing your feedback in different contexts. Chaturvedi Introduced.
Specifically, the machine will present a number of mini case studies to the candidate. The candidate will use the information to answer the question, and the machine will evaluate the answer to determine the career suitable for the candidate's development. For example, according to the company, such tests can tell whether a person is more suitable for marketing or financial work.
Not only that, such machines can see a range of other details. For example, whether users who find mentors for career development advice online have obtained satisfactory answers in the conversation, whether they have maintained an attentive and polite attitude in the communication process, and whether there are emotions in the conversation.
And when the technology is applied to online teaching, it can also assess students' comprehensive capabilities. For example, when a child is thinking intently, the machine can know whether he is thinking about the material he is reading or remembering what he has heard.
Now, almost all experts who study future employment issues will tell you: Machines will leave high-level cognitive materials to humans by completing most of the repetitive work; then from a human perspective, the most important skills in the future are Critical thinking.
AugLi is a technology company located in Gurgaon. They have built a test platform that can analyze the critical thinking ability of children aged 10-16. The basic mechanism and design purpose is to first measure the distinction between the reading age and the physiological age of students, and then push them to read content that is truly suitable for their cognitive ability according to the actual situation of the students.
Most of the test readings are based on current affairs politics, but there are also biographies of some of the big names, such as well-known Indian boxer Mary Kom. After the students have finished reading the article, the platform will ask them to make a brief introduction of the content, which can be audio or text.
We think it's important to find the important content of the article and the ability to quickly absorb and repeat it twice, said Anjali Tiwary, founder and CEO of AugLi. We can see how well you understand the content of the article.
The we in Tiwari's mouth is not human. According to him, the machine algorithm can see the reader's understanding of the content of the article, and this is likely to be achieved through the technical branch of artificial intelligence, natural language processing (NLP).
The specific process is as follows: Take audio repetition as an example, the platform will first transcribe the student's audio content into text, and then the machine algorithm will compare the similarities and differences between the original text and the transcribed text.
The nouns are identified first, then the phrases. If students use words that have never appeared in the original text, the machine will identify them and start searching for related information and analyzing the differences in ideas behind different words. AugLi Union Founder Kamal Kashyap further added, Since the system has been embedded with tens of thousands of relevant texts, it can 'understand' the relationship between different concepts.
Coincidentally, Vendantu is also using big data and natural language processing technology to analyze the learning status of students and find ways to improve learning efficiency. For example, face recognition technology can be used to evaluate students 'learning participation and teachers' teaching effects: Are students staring at the screen or looking elsewhere? Are they smiling or frowning? Do some children blink too often? Why is that student's eyes so wide? What is the time difference between the teacher's question and the student's answer?
Each lesson we capture more than 70 variables. We use the amount of data that other companies can't reach to analyze children's understanding. Vedantu co-founder Vamsi Krishna is full of confidence in his own technical capabilities. Support in these data Next, we can create customized learning content for each student. Real-time learning analysis technology is our strongest patent.
The National Education Policy 2019 Draft sets goals for primary and secondary schools at all levels throughout the country: software must be developed to help teachers evaluate children's teaching fitness; after the periodical assessment is completed, students must be provided with appropriate anti-failures.
The new mechanism will minimize the importance of rote memorization and replace it with a focus on critical and creative thinking, as well as a focus on students' communication and collaboration skills. These are the training programs that meet the needs of talent in the 21st century. Said the ambitious draft policy.