OpenCog Hyperon AGI engine is an open-source framework aiming to facilitate the emergence of Artific...
Our probabilistic logic Bayesian net service, BayesianExpert, makes the manual creation of Bayesian ...
One of the most powerful aspects of Rejuve's AI approach is its ability to effectively integrate patterns found in multiple types of relevant data. The volume of data and computational power needed to solve complex health problems like aging has only recently been made available to research scientists. Much of AI’s impressive recent gains still exists mostly in silos, like healthcare claims, and biomarker, genomic, and imaging studies. Despite this progress, these successful siloed applications only weakly relate to the general issue of aging; and when analyzing a complex system like the human body, siloed artificial intelligence is not enough.
Rejuve seeks to improve this situation by combining these siloed views of the human body to more clearly understand the underlying complex system. Rejuve's analysis tools fuse multiple data channels that will facilitate the interpretation of sparsely overlapping observational data. For example, when a dataset contains patients with measurable variables A and B, as well as patients with measurable variables B and C, Rejuve’s AI can interpret variables ‘A’ and ‘C’ by using the overlapping variable ‘B’. Rejuve’s AI combines specific information and insights to develop a greater high-level understanding of the human body. To be able to combine these silos, Rejuve has to offer a platform that allows for deeply synergistic data integration.
Modelling and understanding the human body is too big a task for any one entity or organization. Leveraging SingularityNET’s open platform of AI, Rejuve is developing methods by which a swarm of researchers can come to consensus on models of increasing complexity, while exploring multiple combinations of models and datasets that make sense to combine together. Metrics of consonance – i.e. formal measurements of the amount of sense that models and data make together – facilitate automated complexification in a decentralized manner. The models submitted by diverse researchers can be combined in an automated way, or via participation of researchers.
Decentralization brings the wisdom of the crowd to the challenge of longevity, and automated models embody the cumulative consensus of researchers. Rejuve’s models of human biology are something tangible and cumulative that the community can work on together, but are larger than reductionist randomized controlled trials and take more variables into account. Concrete models focus the discussion on a specific object, which the hive mind can debug until consensus is reached. Working on common tangible models, and being able to see the consequences of their models for other models, helps researchers find what their models imply, and to think more holistically and systemically than when concentrating on a single variable.
Rejuve’s AI prioritizes causation, leveraging tailored neural techniques combined with probabilistic causal logical inference methods.
Deep neural networks, today’s most popular data modeling method, can learn a system in amazing detail, but at the same time they are poor at making recommendations. Neural networks struggle with suggesting the best treatment from their models of the world, mainly because these models are correlational rather than causal. Correlation is useful for general prediction, but when trying to predict how a treatment will work, principles of causation matter. Rejuve's AI platform prioritizes understanding causation, leveraging appropriately tailored neural techniques combined with probabilistic causal logical inference methods.
Healthcare data analyses must be causal in order to deal with longevity treatments, but should also exhibit generality. This generality helps them be accurate, but moreover allows what is learned in a larger dataset to be applied to another smaller dataset. Not all persons present the same kinds of health data for analyses; some will be power users that regularly take readings of health data from many wearables, have genomic sequencing and medical imaging done, take biomarker tests and try out many supplements, and others will do only a few of those. Rejuve allows people who’ve tracked less data to learn from those who have tracked more. For example, we can interpret genomic and nutritional data from small sets of supercentenarians using knowledge transferred from larger sets of genetic and nutritional data.