Patents

Avata Intelligence holds a number of patents for its sophisticated AI algorithms and solutions

OPTIMIZING ALLOCATION PLANS USING DECOMPOSED OPTIMAL BAYESIAN STACKELBERG SOLVER

US8224681

Describes the algorithm DOBSS, which is the fastest “general purpose Stackelberg game” solver — it can compute the optimal strategy for any domain (e.g. airport security) where the first player (e.g. security forces) has to act first (e.g. execute security policy) and the second player (e.g. adversary) can act second (e.g. attack after observing and learning the strategy of the security forces).

SECURITY ON NETWORKS

US61/792,931 (PROVISIONAL)

This work deals with network based security games, where a defender has to strategically place security measures on the edges of a graph (e.g. in cyber-security or transportation networks) to protect against an adversary who chooses a path through the graph to attack diverse assets. It presents the only and fastest solution to this network based security game problem.

LOCALIZED SHORTEST-PATHS ESTIMATION OF INFLUENCE PROPAGATION FOR MULTIPLE INFLUENCERS

US61/791,273 (PROVISIONAL)

This work presents algorithms for domains where strategic actions carry a ‘contagious’ component in the presence of multiple strategic players. For example, viral marketing on a social network in the presence of many competing agencies, or garnering local support in counterinsurgency domains. This work models strategic actions which have a probabilistic, non-local impact (e.g. strategically influencing a particular user on a social network is required to maximize the propagation of influence on other users of the social network).

COMPETITIVE PLANNING WITH DYNAMIC EXECUTION UNCERTAINTY

US61/790,360 (PROVISIONAL)

This work deals with scheduling in time-critical domains, where it is imperative for a resource to re-plan once a planned schedule is interrupted due to real world constraints. For example, a Sheriff may miss a train because he/she is writing a ticket, and this algorithm will dynamically compute optimal new plans in the presence of such uncertainty.