multivac | MULTIVAC - A DARPA Automating Scientific Knowledge Extraction (ASKE) Artifical Intelligence Explorat | Machine Learning library

multivac is a Python library typically used in Institutions, Learning, Education, Artificial Intelligence, Machine Learning applications. multivac has no bugs, it has no vulnerabilities, it has build file available and it has low support. You can download it from GitHub.

DARPA’s Information Innovation Office’s Automating Scientific Knowledge Extraction (ASKE) program seeks to develop approaches to make it easier for scientists to build, maintain and reason over rich models of complex systems — which could include physical, biological, social, engineered or hybrid systems. By interpreting and exposing scientific knowledge and assumptions in existing model code and documentation, researchers can identify new data and information resources automatically, extracting useful information from these sources, and integrating this useful information into machine-curated expert models for robust modeling. Gallup’s Meta-model Unification Learned Through Inquiry Vectorization and Automated Comprehension (MULTIVAC) effort supports these goals by developing a system that absorbs scientific knowledge — in the form of facts, relationships, models and equations — from a particular domain corpus into a semantic knowledge graph and learns to query that knowledge graph in order to accelerate scientific exploration within the target domain. MULTIVAC consists of an expert query generator trained on a corpus of historical expert queries and tuned dialectically with the use of a Generative Adversarial Network (GAN) architecture. As a prototype system, MULTIVAC focuses on the domain of epidemiological research, and specifically the realm of SIR/SEIR (Susceptible-Infected-Recovered, often with an additional “Exposed” element) compartmental model approaches. It is Gallup’s intent that this system includes a “human-in-the-loop” element, especially during training, to ensure that the system is properly tuned and responsive to the needs and interests of the human researchers it is intended to augment.