for Healthcare
Please inquire about the new function “CKE-LLM (Causal Knowledge Extraction LLM)” that enables the extraction and utilization of known causal relationships using LLM and the use of MSD manuals as data sources.
ABOUT
Accelerates Healthcare Innovation
Causal AI Platform
In this era of diversified lifestyles as well as diversified health risks such as pandemics are diversifying. We at VELDT believe that the various lifelogs that are produced every day hide great hints for changing the lives of individuals with different interests and goals. We will innovate how to utilize the data produced in the real world. We will pursue the following three themes: "Realization of personalized wellness" that is close to one person, "Preventive shift from cure to care" this increases the time that can be active in health and leads to efficiency of medical resources, and lastly "Supporting healthcare innovation from various industries" this is needed because of the care-centered era.
Realize personalized
wellness
Preventive shift
from cure to care
Supporting healthcare
innovation in various industries
Concept
Accelerating Health Care and Wellness Innovations
As diversification accelerates, what is required is personalized disease prevention and treatment, and ingenuity to increase the amount of time one can be healthy and active. xCausal™︎ for Healthcare is a Causal AI platform based on causal relationships that accelerates innovation by combining lifelog data from smartphones, IoT devices, and other devices with data from medical examinations and other social activities.
xCausal use cases
Precision・
Nutrition
Online
Medical Services
R&D DX before
Clinical Trials
Advanced insurance
product development
Smart spaces
optimized for health
Inclusive
Smart Cities
Data-driven
health management
Best Personalized
Conditioning
The Innovation Issue:
just having data is not enough
In our survey of 200 people working for companies that are implementing or planning businesses in wellness and healthcare, approximately 90% answered that they have problems in setting ideas and themes. Of these, 60% recognize that they have data at hand but have not fully utilized it, and about 50% recognize that they rely on their personal experience and intuition.
Discover and virtually validate potential healthcare needs from data
Intermittent and highly reliable data available in the medical and research fields along with daily life log data, which are less reliable but have hints of what happened during that time, are a good complement to each other if the shortcomings are compensated for. By organizing and structuring data, exploring its mechanisms and causal relationships, VELDT provides a mechanism to assist companies and institutions in finding solutions to new problems in virtual verification of hypotheses. We believe that the ability to conduct a virtual validation process prior to conducting a clinical trial or experiment will lead to a more effective experimental design and results.
Service Outline
xCausal is a Web service for companies
and medical institutions used in conjunction with the app
xCausal™︎ uses you'd™, a conditioning AI app that makes physical condition discoveries by utilizing smartphone lifelogs. In addition, Smallytics™ is a powered algorithmic engine that links data from the user at a company or medical institution. The user's data that is within companies and medical institutions can be obtained/ linked and used with permission. The data is anonymized and statistically processed before being used through VELDT's cloud-based dashboards and the construction of causal models. *It is also possible to use your data alone without linking to the application.
※ It is also possible to use your data alone without linking to the application.
you'd™ or
Smallytics™ Onboard Apps
Data Linkage, Visualization,
Causal Inference
Service overview and process flow
xCausal™︎ for Healthcare
To use xCausal, you must first register your company on the web. If you wish to link your lifelogs with the app, you will need to obtain a personal code from the web and distribute it to your company's users. When using the app, the target user enters the assigned personal code into the app and will need to consent to the data linkage. Consent can be given on the app screen after downloading our personal conditioning AI app you'd™ (also available as a white label). You will upload your own data from the web and link it with your personal code (no personal information is required). After the data is linked, you can use the xCausal service in an anonymized and statistical form. It is also possible to upload your own data in csv format for use without linking to the application. The data available via the app is health data in iOS™ Health Care on the user's smartphone or, for Android™, in Google Fit, as well as data about physical condition, ailment symptoms, and habits entered by the user in "you'd™".
Causal discovery and Causal inference from data
xCausal™︎ provides a causal search function to estimate the structure of causal relationships from data. It is also possible to construct causal models with a higher degree of confidence by pre-setting rules and hierarchies of causal relationships that are already known (will enchance the functions) and by manually modifying the estimated structure. On top of that, it is possible to simulate the effect of hypothetically changing values for variables that are available for intervention. Here is a simple example “How would the sleep efficiency change if we increased our daily walking speed?” With xCausal™︎, you can get immediate predictions based on a structural causal model that automatically adjusts for the effects of common causes (confounders) and other factors that are the source of the pseudo-correlation and displays the calculated results.
e.g. “How would the sleep efficiency change if we increased our daily walking speed?”
Virtually verify the effect of intervention
Quick modeling to assist discovery from data
Estimating the structure of a causal graph from a large number of data items requires a long computational process due to the increase in the number of combinations. In xCausal, Smallytics™ is an algorithm developed in-house that calculates and recommends data combinations that are highly relevant in consideration of causality. By using the recommendations to narrow down the number of items, a causal graph can be generated immediately. Because items can be replaced and interventions can be made immediately, trials of causal effects can be repeated quickly, leading to discoveries that have never been made before.
About Linked Apps
Collaborative apps used by users
you'd™
you'd™ is an app that utilizes your daily activity and lifelog data stored in your smartphone to provide useful discoveries for managing your physical condition. The app automatically calculates and displays what is likely to affect you by inputting your own physical condition, symptoms when you are unwell, and habits that you are implementing to manage your physical condition.
you'd™ can be white-labeled and custom-made. You can incorporate it into your own service by changing the UI design, logo, etc., and customizing branding and functions.
Smallytics™ OEM (SDK provided) and
white label instructions
It is possible to use our health data for smartphones and analysis algorithm Smallytics™ as an SDK for existing apps. Smallytics™ has a structure that cooperates with AI on the VELDT's cloud, and we plan to build a mechanism to reflect the intelligence learned from anonymized data and other sources in Smallytics™ on the application. This will provide a mechanism to continuously improve intelligence.
Service Menu
xCausal™︎ for Healthcare
xCausal™︎ offers not only a standard version but also customized services.
xCausal™︎ for Healthcare
- Introduction Workshop / Consulting
- Use of CKE-LLM
(Causal Knowledge Extraction LLM)
(Including MSD manual use) - Customization of analytical environment
- Analysis and report by VELD
- Project use with dedicated AI environment
- Data Check Tool(CSV Checker)
- Causal rule setting function
- Dashboard
- Variable recommendations by Smallytics
- Causal Discovery (Multiple algorithms)
- Causal Inference (Structural Causal Model)
- Mixed use of discrete and continuous values
- Natural Direct Effect (NDE) / Natural Indirect Effect (NIE)
- Robustness Test
- Science Paper search (for healthcare field)
- Consulting Partnerships
Provide consulting services for xCausal utilization - VAR(Value-added resale type) Partnerships
Integration into the user company's services - Custom use of you'd apps
(white label) - Provision of Smallytics™ SDK
to your apps