Nowadays numerous virtual assistants perform outstandingly in their own fields, such as setting navigation on a car, turning on/off electrical appliances in a house, and adjusting schedule on a smartphone. Yet, outside the range, especially during small talks, virtual assistants often have only passable performance. Magi backs up virtual assistants outside their main field with data and processing ability. In short, Magi offers common sense knowledge. Magi KG has real-time knowledge from the whole Internet, offered in the form of structured data that is easy to process, while keeping all the original source for further reference.
Robotic Process Automation
Robotic Process Automation, or RPA, has been widely applied in many organizations. The repetitive jobs in daily operation are taken over by RPA systems, freeing human labor to engage on more challenging tasks. The earlier stage of RPA focus on the more mechanical operation. Combined with AI technology, RPA could benefit much more scenarios, further reducing the time and human capital cost of different tasks. For example, optical character recognition, or OCR, is already serving multiple industries like finance and medical.
Model as a Service, MaaS
For many companies that have little investment into their IT sector, it is challenging to deploy a solution based on artificial intelligence in setting up hardware, building software environment, collecting data, training models, etc. Each part would need the specific specialist for maintenance. Due to cost constraints, flexible development, continuous update, and other reasons, purchasing or outsourcing the entire solution is not always a preferable choice. Besides the traditional API service, Magi directly offers model-related service, including training, deployment, etc.
Inside organizations, knowledge and experience are accumulated over the time and buried in medium like documents and emails. When a new task takes place, even if the past knowledge and experience cannot directly solve the problem, they are usually somehow useful. Through knowledge engineering, the experience could be summarized into a data structure that facilitates convenient consultation, thus improving efficiency and reducing dependency on key experts. Magi has the toolset to assist some key tasks, such as the summarization of documents and information inquiries, in knowledge engineering.
Recommendation system is widely used in many fields. The more traditional algorithm that relies mainly on statistics is relatively naive, while the latest AI-based models have the characteristics of black boxes - results hard to predict, algorithms hard to adjust, process hard to explain, etc. The application of recommendation systems is substantially constrained by shortcomings, especially in scenarios where great importance is attached to reliability, where recommendation systems are given little trust.
Domain-Specific Knowledge Graph
Unstructured data is commonly seen in many industries, including news, press release, media comment, communications between users/clients, etc. Such data provides valuable information for the entire industry. But the collecting of unstructured information typically rely on manual processing, such as summarizing the parameters from media reports into a database. Both the accuracy and timeliness are significantly constrained by the personnel involved. For clients with needs for such data processing, several modules of Magi can improve the efficiency.