Project Overview
Crowd-founding Starting Time: 0 am, Dec 25, 2017
Crowd-founding Ending Time: 0 am, Jan 10, 2018
Basic Information about the Project
Project Description
Overview
The full name of AIDOC is AI Doctor, which is ‘Tian Yi’ in Chinese. It is an international blockchain+AI intelligent medical application development team headquartered in Singapore. The team members of which are from all over the world, and the core members are all top development talents or industry professionals from AI and medical industry, who have several years’ development experience and resource accumulation. in the application of new technology, AI big data in the field of medical industry has many years of development experience and resources accumulated.
With the advent of the era of big data and the continuing improvement of computing power, AI solutions led by deep learning made tremendous progress in the medical industry. As an important branch of artificial intelligence, deep learning is attracting more and more attention and recognition from both academic and industrial circles. Deep learning is a method that uses deep neural network for data modeling. It uses deep neural network to automatically learn the hidden features of data, so as to carry out corresponding classification, regression, segmentation etc., which is shown in the following picture.
Behind deep learning is deep neural network. By simulating human brain neurons, these deep neural networks learn latent features from source data , so as to activate the neurons corresponding to the hidden layers, and finally map to output results. As shown above, as a classic example of deep learning, convolution neural network, based on the local receptive field of human vision, extracts features from visual neuron nodes layer by layer, and then finally integrates local features at high level and map them to the final output. Recurrent neural network can well model time series data, and it can better model the patient’s time based data set, so that information like case and medical record can be better used to give a comprehensive judgement of the patient’s condition.
DAPP in AIDOC takes deep neural network technology as its core, and the improvement of DAPP capability requires a lot of data to train. AIDOC provides online training platform, in which data is authorized to open by the users. It supports a series of mainstream deep learning frameworks like TensorFlow and Caffe. The AI model trainers are unable to take away data, instead, they can only get the conclusion from the AI model. Certain amount of AID tokens, as well as the consumption, are required for the AI model training, and the data provider will be rebated for the consumption of AID tokens.
IDOC,who puts AI model offered by DAPP provider on the AI model chain, and equips independent operating ability, as well as letting the model have the ability to learn incrementally. Compared with the traditional batch learning algorithm, decentralized online incremental learning is a more practical and general method. Generally speaking, it can continuously learn new knowledge from the newly added data around it, and can therefore overcome the drawbacks of redundantly learning previous data, which consumes a lot of time and computing power, and is therefore more consistent with human development process.
Through combining deep learning with blockchain, decentralized online incremental learning is produced. While using blockchain to disguise data and carry out on-chain data sharing, the data itself is increasing. At the same time, the use of online incremental learning method, the existing model is updated quickly and efficiently, which truly produces the prediction of the disease, and lets the model itself grow in the meantime. To sum up, the decentralized online incremental learning system can overcome the disadvantages of the traditional batch learning by using the advantage of the blockchain, that is, fair recording the events on the time line, which really achieves producing the prediction results for patients while learning and upgrading itself.
AI DOCTER is the pioneer in the decentralized intelligent medical field. It applies cutting-edge technological innovations like blockchain and artificial intelligence to medical field, constructing an intelligent medical value chain — AIDOC chain. Participants on the chain are connected, the resources of medical industry are integrated, and the health data value is enhanced, maximizing the value of participants on the chain. With AIDOC, health data is written to blockchain, and bit digital people are cloned in the digital world, so that everyone can have a dedicated AI super doctor. AI DOCTER aims to build an open, equal, safe and intelligent medical on-chain platform, and all participants are able to create and share AIDOC on-chain value. Everyone who contributes data and resources to AIDOC when using the platform will obtain deserving profit.
Project Application
AI DOCTER proposed three core concepts, bit digital person, life bank, and AIDOC.
Bit digital person is a complete digitization of an individual’s vital signs. By wearable devices, smart hardware, sensors and other equipment, the patient’s vital signs are continuously uploaded to the AIDOC chain to form a digitized mapping object of the user ontology in the network, which makes quantitative health management possible. AIDOC chain network can analyze real-time vital signs data, and detect abnormal data in time, and therefore can prevent the risk of unknown diseases.
Life Bank is a place for safely storing medical data and publicly trading in the AIDOC chain. Through encryption technologies like homomorphic encryption, differential privacy (DP), the user’s medical health data is stored safely in the AIDOC decentralized medical network, and is permanently traceable and can never be tampered with. . By uploading data, users can obtain tokens, and their purchases of medical services will consume tokens.
As the “brain” of AIDOC chain, AIDOC integrates intelligent diagnosis modules and API interface, providing users with exclusive private super doctors and complete intelligent medical services. Medp.AI is the first technical service provider of AIDOC chain. The modules currently deployed include image diagnosis, medical bills OCR, diabetic retinopathy screening, ECG monitoring, etc.
Token Structure
Project Benefits
AIDOC has now acquired the strategic technology support from Medp.AI, a team founded in January 2015 by senior scientists from Baidu and Microsoft and well-known professors from top universities, and it is a leading medical AI company in China. The company has successively developed several kinds of AI products, such as intelligent chest CT assisted diagnosis system, intelligent X-ray assisted diagnosis system, ECG intelligent analysis system, medical bills OCR system etc., and has provided services to hundreds of medical institutions. With more than 40 national patents and relevant technology rights, its core technology members have won the best paper award many times, as well as the highest awards in several international evaluations such as FAT, Middlebury, etc. Medp.AI is a high-tech enterprise in both Zhongguancun and the whole country, and it has continuously won the titles of The Best Growing Medical Innovation Company, The Best Artificial Intelligence Start-up Company, and The Best Promising Start-up Company selected by famous media. At the same time, AIDOC and Medp. AI have won hundreds of billions of investments from well-known venture capitals, including the U.S. Redpoint Ventures, New Version Ventures, Union Capital, Aplus Capital, Ciming Health Checkup, as well as investments from famous investment institutions in the blockchain field like Node Capital, Connection Capital, Geek Capital etc.
Roadmap
