Interactive machine learning assistant
In an example embodiment, a method may generate an item layout map for a shelving unit, where the shelving unit stores item types associated with a product category. The method may determine product attributes associated with the product category and divider configurations of the shelving unit, where a divider configuration divides the shelving unit into one or more candidate regions. The method may generate an ontology entry associated with a product attribute and the divider configuration based on the item layout map and the divider configuration of the shelving unit, the ontology entry indicating a regional entropy values of a product attribute for the candidate regions generated by the divider configuration. In some instances, the method may generate a guidance instruction related to the product attribute of the product category based on the ontology entry and provide the guidance instruction to a user.
1. A method comprising:
receiving, by one or more processors, a user input including an input attribute value for a product attribute associated with an ontology entry, the ontology entry indicating one or more regional entropy values;
determining, by the one or more processors, a first region from one or more regions of a shelving unit based on a regional entropy value of the product attribute for the first region;
determining, by the one or more processors, one or more attribute values of the product attribute of one or more item types stored in the first region;
generating, by the one or more processors, a guidance instruction to the first region associated with the input attribute value based on the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value, the guidance instruction providing an instruction to a location of the determined first region; and
providing, by the one or more processors, the guidance instruction to a user via a user interface device.
2. The method of claim 1 , wherein generating the guidance instruction includes:
generating the guidance instruction to navigate the user to a storage position of an item type that satisfies the input attribute value in the shelving unit based on the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value.
3. The method of claim 1 , further comprising:
determining, by the one or more processors, the ontology entry associated with the product attribute of a product category, the product category including the one or more item types placed on the shelving unit, the ontology entry indicating the one or more regional entropy values of the product attribute for the one or more regions of the shelving unit.
4. The method of claim 1 , wherein:
the regional entropy value of the product attribute for a region in the shelving unit indicates a purity level of attribute values of the product attribute in the region; and
determining the first region from the one or more regions of the shelving unit includes:
determining that the first region has the regional entropy value of the product attribute for the first region satisfying a regional entropy value threshold.
5. The method of claim 1 , wherein generating the guidance instruction includes:
performing a comparison between the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value;
determining that the first region stores an item type satisfying the input attribute value based on the comparison; and
responsive to determining that the first region stores the item type satisfying the input attribute value, generating the guidance instruction to the first region of the shelving unit.
6. The method of claim 1 , wherein:
the shelving unit includes one or more space units; and
determining the one or more attribute values of the product attribute of the one or more item types stored in the first region includes:
mapping the first region to an item layout map of the shelving unit, the item layout map indicating the one or more attribute values of one or more product attributes of an item type stored in a space unit of the shelving unit.
7. The method of claim 1 , wherein:
receiving the input attribute value from the user includes receiving the input attribute value in a conversational dialog with the user; and
providing the guidance instruction to the user includes providing the guidance instruction to the user in the conversational dialog with the user.
8. The method of claim 1 , further comprising:
determining one or more divider configurations of the shelving unit, a divider configuration in the one or more divider configurations dividing the shelving unit into one or more candidate regions.
9. The method of claim 8 , further comprising:
generating the ontology entry associated with the product attribute and the divider configuration in the one or more divider configurations based on an item layout map and the divider configuration of the shelving unit, the ontology entry indicating the one or more regional entropy values of the product attribute for the one or more candidate regions generated by the divider configuration.
10. The method of claim 9 , wherein generating the ontology entry associated with the product attribute and the divider configuration includes:
computing the regional entropy value of the product attribute for a candidate region of the one or more candidate regions based on a probability of space units in the candidate region that are associated with an attribute value of the product attribute.
11. The method of claim 9 , further comprising:
computing a first entropy value of the product attribute for the shelving unit based on a probability of space units in the shelving unit that are associated with an attribute value of the product attribute; and
computing a second entropy value of the product attribute for the shelving unit based on the one or more regional entropy values of the product attribute for the one or more candidate regions generated by the divider configuration.
12. The method of claim 11 , further comprising:
computing an information gain ratio associated with the product attribute and the divider configuration based on the first entropy value of the product attribute for the shelving unit and the second entropy value of the product attribute for the shelving unit;
determining that the information gain ratio associated with the product attribute and the divider configuration satisfies an information gain ratio threshold; and
responsive to determining that the information gain ratio associated with the product attribute and the divider configuration satisfies the information gain ratio threshold, generating the ontology entry associated with the product attribute and the divider configuration.
13. A system comprising:
one or more processors; and
a memory storing instructions that, when executed by the one or more processors, cause the system to:
receive a user input including an input attribute value for a product attribute associated with an ontology entry, the ontology entry indicating one or more regional entropy values;
determine a first region from one or more regions of a shelving unit based on a regional entropy value of the product attribute for the first region;
determine one or more attribute values of the product attribute of one or more item types stored in the first region;
generate a guidance instruction to the first region associated with the input attribute value based on the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value; and
provide the guidance instruction to a user.
14. The system of claim 13 , wherein generating the guidance instruction includes:
generating the guidance instruction to navigate the user to a storage position of an item type that satisfies the input attribute value in the shelving unit based on the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value.
15. The system of claim 13 , further comprising:
determining, by the one or more processors, the ontology entry associated with the product attribute of a product category, the product category including the one or more item types placed on the shelving unit, the ontology entry indicating the one or more regional entropy values of the product attribute for the one or more regions of the shelving unit.
16. The system of claim 13 , wherein:
the regional entropy value of the product attribute for a region in the shelving unit indicates a purity level of attribute values of the product attribute in the region; and
determining the first region from the one or more regions of the shelving unit includes:
determining that the first region has the regional entropy value of the product attribute for the first region satisfying a regional entropy value threshold.
17. The system of claim 13 , wherein generating the guidance instruction includes:
performing a comparison between the one or more attribute values of the product attribute of the one or more item types in the first region and the input attribute value;
determining that the first region stores an item type satisfying the input attribute value based on the comparison; and
responsive to determining that the first region stores the item type satisfying the input attribute value, generating the guidance instruction to the first region of the shelving unit.
18. The system of claim 13 , wherein:
the shelving unit includes one or more space units; and
determining the one or more attribute values of the product attribute of the one or more item types stored in the first region includes:
mapping the first region to an item layout map of the shelving unit, the item layout map indicating the one or more attribute values of one or more product attributes of an item type stored in a space unit of the shelving unit.
19. The system of claim 13 , wherein:
receiving the input attribute value from the user includes receiving the input attribute value in a conversational dialog with the user; and
providing the guidance instruction to the user includes providing the guidance instruction to the user in the conversational dialog with the user.
20. A method comprising:
determining an ontology entry including:
computing an information gain ratio associated with a product attribute and a divider configuration based on a probability of space units in a shelving unit being associated with an attribute value of the product attribute for the shelving unit and one or more regional entropy values of the product attribute for the divider configuration for the shelving unit; and
responsive to determining that the information gain ratio associated with the product attribute and the divider configuration satisfies an information gain ratio threshold, generating the ontology entry associated with the product attribute and the divider configuration;
receiving a user input including an input attribute value for the product attribute associated with the ontology entry, the ontology entry indicating the one or more regional entropy values;
determining a region from one or more regions of the shelving unit based on a regional entropy value of the product attribute for the region;
determining one or more attribute values of the product attribute of one or more item types stored in the region;
generating a guidance instruction to the determined region associated with the input attribute value based on the one or more attribute values of the product attribute of the one or more item types in the region and the input attribute value; and
providing the guidance instruction to a user.