Neural network classifiers for block chain data structures

1. A method comprising:

implementing, by a blockchain network including blockchain nodes and in conjunction with a set of distributed machine implemented applications (DApps) communicating with the blockchain nodes, event detection, tracking and management by performing operations including:

detecting at least one block-level event of a set of block-level events recorded in a blockchain ledger provided to the blockchain nodes, wherein a block in the blockchain ledger corresponds to one of the set of block-level events and includes at least an event type data and a set of event data related to the event type data;

classifying, by at least one DApp, of the set of DApps, implemented as a classifier, the at least one block-level event by applying at least the event type data as an input to the classifier to determine a situation state of the at least one block-level event as being one of counterfeit, diverted, stolen, intentional adulteration, unfit for distribution and fraudulent transaction;

triggering a remedial application resident on a server to, in dependence upon the determined situation state, perform at least one of instructing a user to capture an image of a product or package associated with the at least one block-level event, recording a sound associated with the at least one block-level event, scanning a bar code of the product or package associated with the at least one block-level event, and submitting a regulatory report regarding the product or package associated with the at least one block-level event; and

detecting, by at least one DApp, whether any two block-level events of the set of block-level events are separated by a specific period of time and when the two block-level events are not separated by the specific period of time assigning a weighting to both a first temporally occurring block-level event and a second temporally occurring block-level event to indicate that both the first and second temporally occurring block-level events are defective.

2. The method of claim 1 , wherein the classifying by the at least one DApp includes using a classifier implementing statistical learning to select a situation state from a plurality of situation states using an ensemble of neural network classifiers.

3. The method of claim 2 , wherein the ensemble of neural network classifiers includes at least a recurrent neural network (RNN).

4. The method of claim 2 , further including training at least some of the ensemble of neural network classifiers using a training set comprised of a set of ground truth situation states and a labelling set.

5. The method of claim 2 , wherein the ensemble of neural network classifiers is trained with a ground truth dataset comprising pairings of investigative situations with exception events including sensed discrepancies in a physical object or a physical process, and remedial actions including capturing evidence of the sensed discrepancies.

6. The method of claim 1 , further including:

adding, by a triggered application resident on the server, a block to the blockchain ledger based on the situation state to generate a new blockchain ledger that records the situation state as determined; and

outputting a copy of the new blockchain ledger for distribution to at least one other blockchain node of the blockchain nodes.

7. The method of claim 1 , further including triggering an input data collecting application configured to receive input data indicating sensed information about a situation at a site.

8. The method of claim 1 , further including assigning weightings and probabilities to a determined situation state by the at least one DApp of the set of DApps.

9. The method of claim 8 , further including generating a report including the weightings and the probabilities assigned to the determined situation state.

10. The method of claim 1 , further including a DApp detecting whether any two block-level events of the set of block-level events are separated by a specific period of time or threshold and assigning like weightings to a first temporally occurring block-level event and a second temporally occurring block-level event to indicate that each of the first and second temporally occurring block-level events has a probability of being defective.

11. The method of claim 1 , further including:

a DApp detecting that an aggregate number of block-level events added to the blockchain ledger over a time period exceeds a threshold; and

determining that there is a potential shortage of a product in dependence upon the detecting that the aggregate number of block-level events added to the blockchain ledger exceeds the threshold.

12. The method of claim 1 , further including retrieving data from block-level events storing data from exception reports from the blockchain network.

13. The method of claim 1 , further including retrieving data from block-level events storing data from databases external to servers implementing the blockchain network.

14. A non-transitory computer readable medium storing instructions that, when executed by one or more processors, perform a method comprising:

implementing, by a blockchain network including blockchain nodes and in conjunction with a set of distributed machine implemented applications (DApps) communicating with the blockchain nodes, event detection, tracking and management by performing operations including:

detecting at least one block-level event of a set of block-level events recorded in a blockchain ledger provided to the blockchain nodes, wherein a block in the blockchain ledger corresponds to one of the set of block-level events and includes at least an event type data and a set of event data related to the event type data;

classifying, by at least one DApp, of the set of DApps, implemented as a classifier, the at least one block-level event by applying at least the event type data as an input to the classifier to determine a situation state of the at least one block-level event as being one of counterfeit, diverted, stolen, intentional adulteration, unfit for distribution and fraudulent transaction;

triggering a remedial application resident on a server to, in dependence upon the determined situation state, perform at least one of instructing a user to capture an image of a product or package associated with the at least one block-level event, recording a sound associated with the at least one block-level event, scanning a bar code of the product or package associated with the at least one block-level event, and submitting a regulatory report regarding the product or package associated with the at least one block-level event; and

detecting, by at least one DApp, whether any two block-level events of the set of block-level events are separated by a specific period of time and when the two block-level events are not separated by the specific period of time assigning a weighting to both a first temporally occurring block-level event and a second temporally occurring block-level event to indicate that both the first and second temporally occurring block-level events are defective.

15. A system comprising:

an interface to a memory storing a blockchain ledger of block-level events establishing a blockchain network;

a memory storing instructions; and

one or more processors configured to execute the stored instructions to perform a method comprising:

implementing, by the blockchain network including blockchain nodes and in conjunction with a set of distributed machine implemented applications (DApps) communicating with the blockchain nodes, event detection, tracking and management by performing operations including:

detecting at least one block-level event of a set of block-level events recorded in the blockchain ledger provided to the blockchain nodes, wherein a block in the blockchain ledger corresponds to one of the set of block-level events and includes at least an event type data and a set of event data related to the event type data;

classifying, by at least one DApp, of the set of DApps, implemented as a classifier, the at least one block-level event by applying at least the event type data as an input to the classifier to determine a situation state of the at least one block-level event as being one of counterfeit, diverted, stolen, intentional adulteration, unfit for distribution and fraudulent transaction;

triggering a remedial application resident on a server to, in dependence upon the determined situation state, perform at least one of instructing a user to capture an image of a product or package associated with the at least one block-level event, recording a sound associated with the at least one block-level event, scanning a bar code of the product or package associated with the at least one block-level event, and submitting a regulatory report regarding the product or package associated with the at least one block-level event; and

detecting, by at least one DApp, whether any two block-level events of the set of block-level events are separated by a specific period of time and when the two block-level events are not separated by the specific period of time assigning a weighting to both a first temporally occurring block-level event and a second temporally occurring block-level event to indicate that both the first and second temporally occurring block-level events are defective.