Providing property information

Abstract

For providing property information 207, a method receives a property query 201. The method determines whether the property query 201 is a relevant query 203 using a relevance model 251. In response to the relevant query 203, the method translates the property query 201 into a retrieval query 205 with a query translator 253. The method retrieves property information 207 using the retrieval query 205. The method translates the property information 207 into response information 209, wherein the response information 209 comprises a privacy component 231. The method presents the response information 209.

Description

PROVIDING PROPERTY IN FORMATIONBACKGROUN D INFORMATION

[0001] The subject matter disclosed herein relates to providing property information.BRIEF DESCRI PTION

[0002] A method for providing property information is disclosed. The method receives a property query. The method determines whether the property query is a relevant query using a relevance model. In response to the relevant query, the method translates the property query into a retrieval query with a query translator. The method retrieves property information using the retrieval query. The method translates the property information into response information, wherein the response information comprises a privacy component. The method presents the response information. An apparatus and computer program product are also disclosed.BRIEF DESCRI PTION OF DRAWINGS

[0003] A more particular description of the embodiments briefly described above will be rendered by reference to specific embodiments that are illustrated in the appended drawings. Understanding that these drawings depict only some embodiments and are not therefore to be considered to be limiting of scope, the embodiments will be described and explained with additional specificity and detail through the use of the accompanying drawings, in which:

[0004] Figure l is a schematic block diagram illustrating one embodiment of property information system;

[0005] Figure 2A is a schematic block diagram illustrating one embodiment of property data;

[0006] Figure 2B is a schematic block diagram illustrating one embodiment of response information;

[0007] Figure 2C is a schematic block diagram illustrating one embodiment of models;

[0008] Figure 2D is a schematic block diagram illustrating one embodiment of relevance model training data;

[0009] Figure 2E is a schematic block diagram illustrating one embodiment of query translator training data;

[0010] Figure 2F is a schematic block diagram illustrating one embodiment of location model training data;

[0011] Figure 2G is a schematic block diagram illustrating one embodiment of process model training data;

[0012] Figure 3A is a schematic block diagram illustrating one embodiment of a property information process;

[0013] Figure 3B is a schematic block diagram illustrating one embodiment of an information extraction process;

[0014] Figure 4A is a schematic block diagram illustrating one embodiment of a computer;

[0015] Figure 4B is a schematic block diagram illustrating one embodiment of a neural network;

[0016] Figure 5A is a flow chart diagram illustrating one embodiment of a property information method;

[0017] Figure 5B is a flow chart diagram illustrating one embodiment of a model training method;

[0018] Figure 5C is a flow chart diagram illustrating one embodiment of a property listing method;

[0019] Figure 6 is a flow chart diagram illustrating one embodiment of an information extraction method; and

[0020] Figures 7A-J are screen shots illustrating embodiments of a user interfaceDETAILED DESCRIPTION

[0021] Reference throughout this specification to "one embodiment," "an embodiment," or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases "in one embodiment," "in an embodiment," and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment, but mean "one or more but not all embodiments" unless expressly specified otherwise. The terms "including," "comprising," "having," and variations thereof mean "including but not limited to" unless expressly specified otherwise. An enumerated listing of items does not imply that any or all of the items are mutually exclusive and / or mutually inclusive, unless expressly specified otherwise. The terms "a," "an," and "the" also refer to "one or more" unless expressly specified otherwise. The term "and / or" indicates embodiments of one or more of the listed elements, with "A and / or B" indicating embodiments of element A alone, element B alone, or elements A and B taken together.

[0022] Furthermore, the described features, advantages, and characteristics of the embodiments may be combined in any suitable manner. One skilled in the relevant art will recognize that the embodiments may be practiced without one or more of the specific features or advantages of a particular embodiment. In other instances, additional features and advantages may be recognized in certain embodiments that may not be present in all embodiments.

[0023] These features and advantages of the embodiments will become more fully apparent from the following description and appended claims or may be learned by the practice of embodiments as set forth hereinafter. As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method, and / or computer program product. Accordingly, aspects of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as a "circuit," "module," or "system." Furthermore, aspects of the present invention may take the form of a computer program product embodied in one or more computer readable medium(s) having program code embodied thereon.

[0024] The computer readable medium may be a tangible computer readable storage medium storing the program code. The computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, holographic, micromechanical, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing.

[0025] More specific examples of the computer readable storage medium may include but are not limited to a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), an optical storage device, a magnetic storage device, a holographic storage medium, a micromechanical storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, and / or store program code for use by and / or in connection with an instruction execution system, apparatus, or device.

[0026] Program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object- oriented programming language such as MATLAB, Python, Ruby, R, Java, Java Script, Julia,Smalltalk, C++, C sharp, Lisp, Clojure, PHP or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider). The computer program product may be shared, simultaneously serving multiple customers in a flexible, automated fashion.

[0027] The schematic flowchart diagrams and / or schematic block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations. It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the Figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. Although various arrow types and line types may be employed in the flowchart and / or block diagrams, they are understood not to limit the scope of the corresponding embodiments. Indeed, some arrows or other connectors may be used to indicate only an exemplary logical flow of the depicted embodiment.

[0028] The description of elements in each figure may refer to elements of proceeding figures. Like numbers refer to like elements in all figures, including alternate embodiments of like elements.

[0029] Figure l is a schematic block diagram illustrating one embodiment of property information system 100. A user may pose a query to the property information system 100, along with any additional graphical user interface (GUI) input. In response, the system 100 produces an array of solution recommendations that are optimally presented via a range of graphical and functional fidelity. The ability for the user to perpetuate an answer seeking context forward, by either a user's own organic inspiration, the system's suggestion, or both, is the defining characteristic of embodiments.

[0030] Information about a property 120 may reside in a searchable index 130 and / or an information storage system 125. The information about the property 120 may have both numeric properties and literate properties. In addition, the information may be organized in a wide variety of formats. For example, each real estate broker may organize information foreach property 120 differently, resulting in numerous incompatible formats. In addition, owners may track information in different ways so that information may have a variety of labels and base units. For example, property revenue may be labeled as monthly rent, annual lease revenue, and the like, each with different numeric values for equivalent levels of revenue. In addition, potential buyers may also seek information about the property 120 that is expressed in different formats, labels, and base units.

[0031] There is a need to organize information about properties 120 so that desired information can be provided rapidly, regardless of the original format of the information and / or the desired format of the information. The property information system 100 organizes and provides information regarding multiple properties 120 as will be described hereafter. In the depicted embodiment, the property information system 100 includes the information storage system 125, the searchable index 130, a server 105, a network 115, and a computer 110. The information storage system 125 may comprise at least one data storage drive such as hard disk drives. The information storage system 125 may store information for a plurality of properties 120. The information may be submitted to the property information system 100. In addition, the information may be scraped from websites, emails, texts, printed material, and the like.

[0032] The server 105 may extract, translate, process, and organize the information in the searchable index 130. A user may query the searchable index 130 and / or the information storage system 125 from a computer 110 via a network 115 such as the Internet. Because of the processing of the information, searches for relevant information are faster and more efficient, increasing the efficiency of the property information system 100.

[0033] Figure 2A is a schematic block diagram illustrating one embodiment of property data 200. The property data 200 may be organized as a data structure in a memory. In the depicted embodiment, the property data 200 includes a property query 201, a relevant query 203, a retrieval query 205, property information 207, response information 209, description text 211, evaluation data 213, a rent roll 215, unnormalized table information 217, a normalized rent table 219, a property location 221, geocoded coordinates 223, an error response 225, rent information 227, page text 229, target titles 241, a normalized table form 243, a property listing 245, a processed listing 247, a responsible party 249, a listing permission 231, and property attributes.

[0034] The property query 201 may be submitted to the property information system 100. The property query 201 may request information regarding at least one property 120. In one embodiment, the property query 201 request information for a plurality of properties 120.

[0035] The property query 201 may request aggregate response information 209. The property query 201 may be for one or more elements of evaluation information 213 that are within a specified range. In addition, the property query 201 may be for the top k properties 120 with specified elements of the evaluation information 213 that are within the specified range. For example, the property query 201 may be for the top 10 properties 120 with the sales price of between five and ten million dollars and annual net revenue of at least 500,000.

[0036] The property query 201 may specify an area such as a city, county, postal code, neighborhood, distance from a specified location, or combinations thereof. The property query 201 may include a property type such as commercial building, multi-family building, medical building, and the like.

[0037] The property query 201 may specify ranges of evaluation data 213. For example, the property query may specify an occupancy rate of 90 to 100 percent.

[0038] The relevant query 203 is a property query 201 requesting information that is relevant to the property information system 100. For example, a property query 201 requesting information about properties 120 for sale is a relevant query 203. A property query 201 requesting locations of Thai restaurants may not be a relevant query 203. The relevant query 203 may include all elements of the property query 201.

[0039] The retrieval query 205 is a translation of the property query 201 and / or relevant query 203 for accessing information storage system 125 and / or searchable index 130.

[0040] The property information 207 comprises information about properties 120. The property information 207 may include aggregate information for multiple properties 120, information for an individual property 120, discrete information for units of a property 120, and the like.

[0041] The response information 209 is presented to a user in response to the retrieval query 205. The response information 209 may be translated from the property information 207. In one embodiment, the response information 209 comprises property attributes 233 and / or a normalized rent table 219.

[0042] The descriptive text 211 may be extracted from the property information 207. The description text 211 may be entered manually, scanned in using optical character recognition (OCR), scraped from a website, parsed from an email, parsed from a text message, and the like.

[0043] The evaluation data 213 may be identified from the property information 207 and / or description text 211. Table 1 illustrates one embodiment of elements that may be included singly or in combination in the evaluation data 213.

[0044] Table 1

[0045] The rent roll 215 may describe the sources and amounts of revenues for a property 120. The rent roll 215 may be described by multiple labels, employ multiple base units, and be organized in multiple formats. In one embodiment, the rent roll 215 comprises a tenant, a lease start date, a lease end date, and at least two of a total area, a monthly rent amount, an annual rent amount, a monthly rent per area, and an annual per area.

[0046] The evaluation data 213 may be extracted as unnormalized table information 217. The unnormalized table information 217 may be organized in multiple formats employing multiple labels and multiple base units.

[0047] The normalized rent table 219 may be generated from the unnormalized table information 217 as will be described hereafter.

[0048] The property location 221 may specify a location for a property 120. The property location 221 may be a discrete address. In addition, the property location 221 may be a unit within a discrete address. In one embodiment, the property location 221 is an area such as a government jurisdiction, a neighborhood, a postal code, an area within a given distance of the specified location, and the like. In one embodiment, the geocoded coordinates 223 specify the property location 221.

[0049] The error response 225 may be presented to a user in response to an irrelevant property query 201 as will be described hereafter.

[0050] The rent information 227 may comprise the rent roll 215. The rent information 227 may describe aspects of income for a property 120 under multiple labels, with multiple base units, and with multiple formats. Table 2 illustrates one embodiment of elements comprising the rent information 227, either singly or in combination.

[0052] The page text 229 may comprise a page of property information 207. The page text 229 may be parsed from a webpage, parsed from an email or text, and / or scanned from a physical page. The target titles 241 may be desired targets and / or labels for the response information 209. In one embodiment, the target titles 241 are equivalent to the evaluation data 213.

[0053] The normalized table form 243 may be a template for the response information 209 and / or the normalized rent table 219. In one embodiment, one of a plurality of normalized table forms 243 may be selected.

[0054] The property attributes 233 describe attributes of a property. The property attributes 233 may include the evaluation data 213 and / or rent information 227. In addition, the property attributes 233 may include one or more elements of table 3.

[0056] Figure 2B is a schematic block diagram illustrating one embodiment of the response information 209. In the depicted embodiment, the response information 209 includes a privacy component 231, a property map 233, and / or an itemized list 235.

[0057] The privacy component 231 may grant access to a given user for the given user's property information 207 and public property information 207. The privacy component 231 may specify whether the response information 209 may be shared. In one embodiment, the privacy component 231 specifies individuals and / or organizations with which the response information 209 may be shared. The privacy component 231 may restrict the forwarding of the response information 209. In addition, the privacy component 231 may restrict the printing of the response information 209.

[0058] The property map 233 may include the property location 221 and / or geocoded coordinates 223. In one embodiment, the property map 233 shows a given property 120 as a top view schematic and / or a top view photo. In a certain embodiment, the property map 233 shows given property 120 along with adjacent properties 120.

[0059] The itemized list 235 may include at least one element from one or more of the property data 200, evaluation data 213, and / or rent information 227.

[0060] The property listing 245 offers a property for sale and / or lease. The property listing 245 may be presented by a responsible party 249 that is third-party and / or outside party. The property listing 245 may be a web page, a database entry, an email, a text message, and the like. The listing permission 231 may be received from the responsible party 249 and grant permission to present the processed listing 247. The processed listing 247 may be a property listing 245 that is processed for presentation as will be described hereafter.

[0061] Figure 2C is a schematic block diagram illustrating one embodiment of models 250. The models 250 may be employed to identify, evaluate, translate, and / or process the property information 207 into response information 209. In the depicted embodiment, the models 250 include a relevance model 251, a query translator 253, a location model 257, and a process model 259. The models 250 may be fine-tuned large language models (LLM). The models 250 may each comprise at least one of a neural network model, a deep learning model, a reinforcement learning model, an unsupervised learning model, a supervisedlearning model, a random forest model, a text to vector model, a transformer model, and the like.

[0062] Figure 2D is a schematic block diagram illustrating one embodiment of relevance model training data 270. The relevance model training data 270 may be used to train the relevance model 251. In the depicted embodiment, the relevance model training data 270 includes a plurality of user queries 271 and corresponding prompts 261 and responses 273. The prompt 261 may be a system prompt 261. The prompt 261 may comprise a varying amount of detail. In one embodiment, the size and / or detail of the prompt 261 may be decreased to increase the accuracy of the relevance model 251. Alternatively, increasing the size and / or detail of the prompt 261 may decrease the accuracy of the relevance model 251. In a certain embodiment, no prompt 261 is used.

[0063] Each response 273 may indicate whether the corresponding user query 271 is relevant or irrelevant. The response 273 may be autonomously labeled and / or corrected. In one embodiment, the user query 271 is relevant if response information 209 generated by the property information system 100 answers the user query 271. In addition, our response 273 may indicate that the corresponding user query 271 is irrelevant if the response information 209 generated by the property information system 100 does not answer the user query 271 to a user satisfaction. The responses 273 may be gathered along with the corresponding user query 271 from the user survey. Alternatively, the user query 271 and corresponding response 273 may be algorithmically generated based on the user interaction with the property information system 100.

[0064] In one embodiment, the training data 270 includes training parameters 275. The training parameters 275 may be specified to train the relevance model 251.

[0065] The relevance model training data 270 may include model targets 277. The model targets 277 are used to determine if the relevance model 251 is satisfactorily trained.

[0066] Figure 2E is a schematic block diagram illustrating one embodiment of query translator training data 280. The query translator training data 280 may be used to train the query translator 253. In one embodiment, the query translator 253 is an LLM. In the depicted embodiment, the query translator training data 280 includes a plurality of user queries 271 and corresponding translated user queries 281. The translated user queries 281 may be elastic user queries. The elastic translated user queries 281 may be used to access an elastic search database and / or an Opensearch® database. In a certain embodiment, the translated user query 281 is translated from the corresponding user query 271. The translated user query 281 may be stateless, scalable, and / or cache less. In one embodiment,the user query 271 comprises the property query 201 and a corresponding prompt that is stored with the property query 201. The translated user query 281 is used to query the information storage system 125. The property query 201, corresponding prompt, and translated user query 281 may be evaluated for correctness and / or to evaluate an effectiveness of the query translator 253.

[0067] In one embodiment, the query translator training data 280 includes the training parameters 275. The training parameters 275 may be specified to train the query translator 253.

[0068] The training data 280 may include model targets 277. The model targets 277 are used to determine if the query translator 253 is satisfactorily trained.

[0069] Figure 2F is a schematic block diagram illustrating one embodiment of location model training data 290. The location model training data 290 may be used to train the location model 257. In the depicted embodiment, the location model training data 290 comprises pairs of user queries 251 and corresponding property locations 221. In one embodiment, the property locations 221 comprise geocoded coordinates 223.

[0070] In one embodiment, the location model training data 290 includes training parameters 275. The training parameters 275 may be specified to train the location model 257.

[0071] The location model training data 290 may include model targets 277. The model targets 277 are used to determine if the location model 257 is satisfactorily trained.

[0072] Figure 2G is a schematic block diagram illustrating one embodiment of process model training data 260. The process model training data 260 may be used to train the process model 259. In the depicted embodiment, the process model training data 260 includes a plurality of prompts 261 and corresponding responses 263. A prompt 261 may include an example of a property listing 245. The corresponding response 263 may indicate whether the property listing 245 is currently available. In addition, the response 263 may indicate whether the property listing 245 is desirable for processing.

[0073] In one embodiment, the response 263 may illustrate a formatting for the corresponding prompt 261. For example, if the property listing prompt 261 lists property information 207, the prompt 261 may illustrate the same property information 207 formatted as response information 209.

[0074] In one embodiment, the process model training data 240 includes training parameters 275. The training parameters 275 may be specified to train the process model 259.

[0075] The process model training data 240 may include model targets 277. The model targets il are used to determine if the process model 259 is satisfactorily trained.

[0076] Figure 3A is a schematic block diagram illustrating one embodiment of a property information process 300. The property information process 300 may be performed by the server 105 and / or the computer 110. The property information process 300 receives a property query 201 and presents response information 209 in response to the property query 201. The property query 201 may be received via the computer 110. For example, the user may input the property query 201 via a mobile telephone computer 110.

[0077] The relevance model 251 determines whether the property query 201 is relevant 302. The property query 201 may be relevant 302 if the relevance model 251 determines that the property management system 100 can provide the requested response information 209. If the relevance model 251 determines that the property query 201 is not relevant 302, the property query 201 is stopped.

[0078] The query translator 253 translates the property query 201 and / or relevant query 203 into a retrieval query 205. The retrieval query 205 is used to retrieve property information 207 from the information storage system 125.

[0079] If the retrieval query 205 includes a property location 221, the location model 257 specifies the property location 221 for the retrieval query 205. If the location model 257 determines the property location 221, a geocode API 210 may encode the property location 221 as geocoded coordinates 223. The retrieval query 205 may include the property location 221 and / or geocoded coordinates 223. In one embodiment, the property location 221 and / or geocoded coordinates 223 are injected into the retrieval query 205 to improve the accuracy of the retrieval query 205.

[0080] In one embodiment, the retrieval query 205 comprises a concatenation of the property query 201 and / or relevant query 203. In addition, the retrieval query 205 may comprise a concatenation of the property query 201, the relevant query 203, the property location 221, and / or the geocoded coordinates 223.

[0081] The information storage system 125 may comprise a database that is queried by a structured language. In addition, the information storage system 125 may comprise a relational database such as a Structure Query Language (SQL) database that is queried with a translated elastic user query 281. In addition, the information storage system 125 may comprise an elastic search database and / or an Opensearch® database. The information storage system 126 may support numeric retrieval techniques and / or text retrieval techniques. The information storage system 125 may support Term Frequency / lnverseDocument Frequency (TT / IDF), BM25 retrieval, vector-based text encoding, ranked cosine similarity, word2vec, latent Dirichlet allocation (LDA), SentenceTransformers, and / or Retrieval Augmented Generation (RAG).

[0082] The searchable index 130 may be used to retrieve the property information 207. The property information 207 may be retrieved if a success criteria is satisfied. The success criteria may be that all elements of the retrieval query 205 are satisfied. Alternatively, the success criteria may be that a specified threshold of elements of the retrieval query 205 are satisfied. The property information 207 is translated to response information 209 and presented to the user at the computer 110.

[0083] Figure 3B is a schematic block diagram illustrating one embodiment of an information extraction process 350. The information extraction process 350 extracts description text 211 and evaluation data 213 from property information 207 and stores the information in the searchable index 130 and / or information storage system 125. The information extraction process 350 may be performed by the server 105 and / or the computer 110.

[0084] The process 350 extracts 351 page text 229. The page text 229 may be property information 207. The page text 229 may be extracted from a webpage, from an email, from a text message, from a printed page, from a video, from an audio file, and the like. The process 350 determines 353 if the page text 229 includes evaluation data 213. The page text 229 may have evaluation data 213 if the page text 229 includes any of the elements of Table 1. If the page text 229 has no evaluation data 213, the process 350 ends 367 table extraction and / or information extraction.

[0085] If the page data 229 includes evaluation data 213, the process 350 extracts 355 the evaluation data 213. In one embodiment, a label corresponding to an evaluation data element and subsequent information until a next label is extracted.

[0086] The process 350 stores 357 the extracted evaluation data 213 as unnormalized table information 217. The unnormalized table information 217 is the extracted evaluation data 213 that hasn't been organized for the normalized rent table 219. In one embodiment, the unnormalized table information 217 is stored in the information storage system 125 and / or searchable index 130.

[0087] The process 350 compares 359 column titles and / or labels in the unnormalized table information 217 to the target titles 241. If the target titles 241 are not present 361 in the unnormalized table information 217, the process 350 ends 367 table extraction. If the target titles 241 are present 361 in the unnormalized table information 217, the process 350constructs 363 the normalized table form 243 from the unnormalized table information 217. In one embodiment, column titles in the unnormalized table information 217 and corresponding information are inserted into the normalized table form 243 at the corresponding target titles 241.

[0088] The process 350 further compares 365 the column titles of the unnormalized table information 217 to the target titles 241 and if the column titles are equivalent to the target titles 241, the target titles 241 and / or the normalized table form 243 is added to the searchable index 130.

[0089] Figure 4A is a schematic block diagram illustrating one embodiment of a computer 110. The computer 110 may be embodied in the computer 110 and / or the server 105 of Figure 1. In the depicted embodiment, the computer includes a processor 405, a memory 410, and communication hardware 415. The memory 410 may store code and / or data. The processor 405 may execute the code and process the data. The communication hardware 415 may communicate with other devices such as the information storage system 125 and / or the searchable index 130. One or more of the relevance model 251, the query translator 253, and / or the location model 257 may be embodied in whole or in part in the computer 110.

[0090] Figure 4B is a schematic block diagram illustrating one embodiment of a neural network 475. One or more of the relevance model 251, the query translator 253, and / or the location model 257 may be embodied in whole or in part in the neural network 147. The neural network 475 includes input neurons 450, hidden neurons 455, and output neurons 460. The neural network 475 may be trained by specifying input values for the input neurons 450 and output values for the output neurons 460 and adjusting parameters for the hidden neurons 455 accordingly. In addition, the neural network 475 may generate output values at the output neurons 460 in response to input values at the input neurons 450 based on the training of the hidden neurons 455. Any number and / or arrangement of neurons 454 / 455 / 460 may be employed.

[0091] Figure 5A is a flow chart diagram illustrating one embodiment of a property information method 500. The method 500 presents response information 209 in response to the property query 201. The method 500 may be performed by the computer 110, the processor 105, and / or the neural network 475, singly or in combination.

[0092] The method 500 starts, and in one embodiment the method 500 receives 501 the property query 201. The property query 201 may be received 501 at any computer 110 or by the server 105.

[0093] The method 500 determines 503 whether the property query 201 is a relevant query203. The relevance of the property query 201 may be determined using the relevance model251. If the property query 201 is not a relevant query 203, the method 500 ends. In one embodiment, the error response 225 is presented in response to an irrelevant property query 201.

[0094] If the property query 201 is a relevant query 203, the method 500 translates 505 the property query 201 and / or relevant query 203 into a retrieval query 205. The query translator 253 may translate 505 the property query 201 into the retrieval query 205.

[0095] If the property query 201 and / or relevant query 203 includes a location 221, the method 500 may extract 509 the property location 221 from the property information 207. The property location 221 may be extracted 509 from the property information 207 using the location model 257 and / or geocoded coordinates 223. In one embodiment, the method translates 511 the property location 221 into geocoded coordinates 223. The property location 221 and / or geocoded coordinates 223 may be appended to the retrieval query 203.

[0096] The method 500 retrieves 507 property information 207 using the retrieval query 203. The method 500 may retrieve 507 the property information 207 if there is no location 221 in the property query 201 and / or relevant query 203. The property information 207 may be retrieved 507 from the information storage system 125. In addition, the property information 207 may be retrieved 507 from the information storage system 125 using the searchable index 130 to identify the property information 207 in the information storage system 125.

[0097] In one embodiment, the method 500 translates 513 the property information 207 into response information 209. An embodiment of translating 513 the property information 207 is shown in Figure 6. The method 500 further presents 515 the response information 209 and the method 500 ends.

[0098] Figure 5B is a flow chart diagram illustrating one embodiment of a model training method 550. The model training method 550 may be employed to train the reference model 251, the query translator 253, and / or the location model 257. The model training method 550 may be performed by the computer 110, the processor 405, and / or the neural network 475, singly and / or in combination.

[0099] The method 550 starts and in one embodiment, the method 550 generates 551 model training data 270 / 280 / 290. The model training data 270 / 280 / 290 may be generated from a data set of historic property queries 201. In addition, the model training dataY1270 / 280 / 290 may be added to from subsequent property queries 201. In one embodiment, the model training data 270 / 280 / 290 is added to in real time with property queries 201.

[0100] The method 550 may set aside 553 a portion of the model training data 270 as test data. The test data will not be used to train the model 251 / 253 / 257.

[0101] The method 550 may specify 555 the training parameters 275. Elements of the property data 200 may be training parameters 275. The model 251 / 253 / 257 is trained 557 using the model training data 270 / 280 / 290 in accordance with the training parameters 275.

[0102] The method 550 generates 559 a prediction from the model 251 / 253 / 257 with the test data. The prediction may be the response 273, the translated user query 281, and / or the property location 221. The method 550 determines 561 whether the prediction satisfies the target model 277. If the prediction does not satisfy the target model Tin the training parameters 275 are modified 555 and the model 251 / 253 / 257 is again trained 557. If the prediction satisfies the target model n , the trained model 251 / 253 / 257 is employed 563 and the method 550 ends.

[0103] Figure 5C is a flow chart diagram illustrating one embodiment of a property listing method 570. The method 570 finds and presents outside property listings. The property listing method 570 may be performed by the computer 110, the processor 405, and / or the neural network 475, singly and / or in combination.

[0104] The method 570 may find 571 an outside property listing 245. In one embodiment, the process model 259 finds 571 the property listing 245 by scanning media such as webpages, databases, emails, text messages, and the like.

[0105] The method 570 parses 573 the property listing 245. In one embodiment, a responsible party 249 is parsed 573 from the property listing 245. In addition, the property information 207 is parsed 573 from the property listing 245.

[0106] The method 570 may obtain 575 the listing permission 231 from the responsible party 249. In one embodiment, the method 570 contacts the responsible party 249 and records the responsible party's response. If the response is positive, the listing permission 231 is recorded. If the response is negative, the method 570 ends.

[0107] If the response is positive, the outside property listing 245 is processed 577 and presented 579 as a processed listing 247 as described in Figure 5A. The method 570 determines 581 whether the outside property listing 245 is active. If the property listing 245 is active, the method 570 continues to present 579 the processed listing 247.

[0108] The method 570 may determine 581 that the outside property listing 245 is active in response to the outside property listing 245 still being available in an offering where theoutside property listing 245 was originally found. In one embodiment, the outside property listing 245 is active if the outside property listing 245 has been updated within a specified time period. In a certain embodiment, the outside property listing 245 is active if the outside property listing 245 is found at a second media such as in the second database listing.

[0109] If the method 570 determines 581 that the outside property listing 245 is not active, the method 570 removes 583 the processed listing 247 and the method 570 ends.

[0110] Figure 6 is a flow chart diagram illustrating one embodiment of an information extraction method 600. The method 600 extracts description text 211 from property information 207. The method 600 may be performed by the computer 110, the processor 105, the neural network 475, or combinations thereof.

[0111] The method 600 extracts 601 description text 211 from property information 207. In one embodiment, the description text 211 is extracted from the unnormalized table information 217.

[0112] The method 600 may identify 603 the rent roll 215. In addition, the method 600 may identify 603 the rent information 227. In one embodiment, the rent roll 215 is identified 603 by identifying elements of the rent information 227. For example, the tenant name and the monthly rent amount of the rent information 227 may be identified as the rent roll 215.

[0113] The method 600 extracts 605 the rent roll 215 and / or rent information 227 from the description text 211. In one embodiment, a text extractor such as AWS Textract® and / or Google Document Al is used to extract 605 the rent roll 215 and / or rent information 227. In addition, the method 600 may employ the process model 259 to extract 605 the rent roll 215.

[0114] The method 600 organizes 607 the response information 209. The response information 209 may include the property attributes 233 and / or the normalized rent table 219 from the rent roll 215 and rent information 227 extracted from the description text 211. In one embodiment, the rent roll 215 and / or extracted rent information 227 is modified to conform to the normalized rent table 219 and / or normalized table form 243. The method 600 may employ the process model 259 to generate the normalized rent table 219. The method 600 presents 609 the response information 209 and the method 600 ends. The response information 209 may be presented 609 in a specified format.

[0115] Figures 7A-J illustrate embodiments of user interfaces. The property information system 100 is predicated on a collaborative feedback loop that can take place between a user and the system 100. This technology allows each point of interaction to be expanded upon to provide an array of things beneficial to each side of the collaboration.

[0116] For a user, collocated GUI affordances that aid in the adoption of this new technology alongside increased efficiency and specificity of expression for their query. The system 100 also allows the dynamic serving of the most helpful response information 209 including content and functional affordances at a greatly increased level of granularity across the user base.

[0117] The entirety of the user interface being integrated into the system 100 to allow a much richer presentation of response information 209 alongside an equally rich capability of subsequent input and learnings. An ever growing library of GUI components constructed in real-time to optimally deliver the system's response information 209 while providing a range of affordances for feedback and follow-up queries is robust and extendable.

[0118] The net result is an increased efficiency and accuracy of expression for the user and the system 100. The user's expression is unimpeded when it is either clear to them what they want to express or how they want to express it. The system 100 is more capable of expressing a range of response information 209 and functionality based on the user's input and or the system's user interfaces. The end result is a more dynamic and modular approach to query and answer expression that is inherently geared towards incremental improvement and learning rather than explicitly designed end-to-end experiences. These include a blend of text and numerical response information 209 and user configurable filters.

[0119] In Figure 7A, an embodiment of a search interface 700 is shown. A user may enter a natural language query 701. In addition, the search interface 700 presents dynamic suggested filters 703. The suggested filters 703 are contextually updated as the natural language query 701 is entered.

[0120] Figure 73 shows an alternate embodiment of the search interface 700 with suggested filters 703 presented by filter category 7071-c. In addition, the search interface 700 presents search prompts 709. The search prompts 709 may be generated based on the natural language query 701 and / or a suggested filter 703. The search prompts 709 may be contextually updated as the natural language query 701 is entered and / or as a suggested filter 703 is selected.

[0121] Figure 7C illustrates one embodiment of a listing display 710. The listing display 710 may be presented in response to the natural language query 701. In the depicted embodiment, the listing display 710 includes suggested filters 703, property listings 711, a listing map 713, a feedback query 715, and the natural language query 701. In one embodiment, each property listing 711 comprises at least one element of response information 209. The listing map 713 may show a location of property listings 711. Thefeedback query 715 may illicit feedback from the user on the relevance of the property listings 711. The natural language query 701 may be used to modify a previous natural language query 701.

[0122] Figure 7D illustrates one embodiment of advanced filters 717. The advanced filters 717 may further refine a natural language query 701. The advanced filters 717 may include one or more elements of Table 4.

[0123] Table 4

[0124] Figure 7E shows a listing display 710 with one property listing 711 selected. Figure 7F shows the listing display 710 of Figure 7E with multiple property listings 711 selected.

[0125] Figure 7G shows a collaborative interaction loop between response information 209.In the depicted embodiment, a listing display 710, a property listing 711, and a propertydocument 719 are shown. A user may iteratively navigate between the response information 209.

[0126] Figure 7H shows one embodiment of suggestion affordances 721. In the depicted embodiment, a dynamic suggestion affordance 721 is presented in response to a listing display 710 and / or listing map 713. The suggestion affordance 721 may be presented in response to a navigation history, a market segmentation, a search system state, and the like.

[0127] Figure 71 illustrates one embodiment of a user interface affordance 723. In the depicted embodiment, the user interface affordance 723 may be used to add map overlays to the listing map 713 such as for foot traffic, population, median income, points of interest, and the like.

[0128] Figure 7J illustrates one embodiment of data source selection 725. In the depicted embodiment, a user may select between sources of property information 207.

[0129] There is a need to organize property information 200 about a plurality of properties 120 so that the desired response information 209 can be provided rapidly, regardless of the original format of the property information 207 and / or the desired format of the response information 209 when presented. The present invention identifies, parses, translates, organizes and provides response information 207 regarding multiple properties 120, providing the needed information as can best be utilized by a user.

[0130] This description uses examples to disclose the invention and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the invention is defined by the claims and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements with insubstantial differences from the literal language of the claims.

[0131] The embodiments may also comprise:

[0132] A method comprising: extracting description text from property information; in response to identifying evaluation data in the property information, extracting the evaluation data as unnormalized table information, wherein the evaluation data comprises at least one of a rent roll, a comparable properties table, a lease comparison, a rent comparison, a sales comparison, a sales comparison, comparisons, financial statements, financial projections, unit information, and demographic information;organizing the description text and evaluation data into response information, the response information comprising property attributes and a normalized rent table generated from the unnormalized table information; and presenting the response information in a specified format.

[0133] The method of claim 1, wherein the rent roll is identified in response to rent information comprising at least one of a tenant name, a lease start date, a lease end date, a vacancy, a monthly rent amount, an annual rent amount, a rent due date, a free date, a lease expiration date, a lease term, a total area, a monthly rent per area, and an annual rent per area.

[0134] The method of claim 1, wherein the normalized rent table comprises rent information comprising at least one of open spaces, occupancy rate, under market rents, over market rents, and lease expiration dates within a specified end period.

[0135] A method comprising: obtaining a listing permission for an outside property listing; process the outside property listing to generate a processed listing; determining whether the outside property listing is active; and removing the processed listing in response to the outside property listing being inactive.

[0136] The method of claim 4, wherein the outside property listing is active if the outside property listing is available in an offering where the outside property listing was originally found.

[0137] The method of claim 4, wherein the outside property listing is active if the outside property listing has been updated within a specified time period.

[0138] The method of claim 4, wherein the outside property listing is active if the outside property listing is found in a second media.

[0139] The method of claim 4, the method further comprising: finding the outside property listing; and parsing the outside property listing.

[0140] A method comprising: presenting response information; presenting an affordance that suggests additional response information based on the presented response information; and presenting the additional response information in response to a user selection.

Claims

AMENDED CLAIMS received by the International Bureau on 07 August 2024 (07.08.2024)Claims1. A method comprising: extracting, by use of a processor, description text from property information using a text extractor; identifying property information from the description text using a process model, wherein the property information comprises a rent and an area and the process model is trained on a training data set of lease revenue, rentable square feet, usable square feet, and / or vacant square feet; organizing response information from the property information to conform to a normalized rent table; and presentingthe response information.

2. The method of claim 1 , the method further comprising: receiving, by use of a processor, a property value; determiningwhetherthe property query is a relevant query using a relevance model; in response to the relevant query, translating the property query into a retrieval query with a query translator; wherein the relevance model is a fine-tuned large language model (LLM) trained on user queries labeled as relevant or irrelevant and wherein an error response is presented in response to an irrelevant property queries.

3. The method of claim 2, wherein the query translator is a fine-tuned LLM trained on pairs of user queries and translated user queries.

4. The method of claim 2, wherein the relevance model and / or the query translator comprises at least one of a neural network model, a deep learning model, a reinforcement learning model, an unsupervised learning model, a supervised learning model, a random forest model, text to vector model, and a transformer model.

5. The method of claim 2, the method further comprising: extracting a property location from the property query using a location model ; and translating the property location into geocoded coordinates.

6. The method of claim 5, wherein the location model is trained on pairs of user queries and property locations.

7. The method of claim 2, wherein the privacy component grants access to a given user for the given user’s property information and public property information.

8. The method of claim 2, wherein the response information comprises at least one of a property map and an itemized list.

9. The method of claim 2, wherein the relevance model is modified for a subsequent property query based on the property query.

10. The method of claim 2, wherein the property query is a natural language query.11 . The method of claim 2, wherein the property query requests aggregated property information for at least two properties.

12. An apparatus comprising: a processor executing code stored in a memory to perform: extracting description text from property information using a text extractor; identifying property information from the description text using a process model, wherein the property information comprises a rent and an area and the process model is trained on a training data set of lease revenue, rentable square feet, usable square feet, and / or vacant square feet; organizing response information from the property information to conform to a normalized rent table; and presentingthe response information.

13. The apparatus of claim 12, the processor further: receiving a property value; determiningwhetherthe property query is a relevant query using a relevance model; in response to the relevant query, translating the property query into a retrieval query with a query translator; wherein the relevance model is a fine-tuned large language model (LLM) trained on user queries labeled with responses as relevant or irrelevant and wherein an error response is presented in response to an irrelevant property queries.

14. The apparatus of claim 13, wherein the query translator is a fine-tuned LLM trained on pairs of user queries and translated user queries.

15. The apparatus of claim 13, wherein the relevance model and / or the query translator comprises at least one of a neural network model, a deep learning model, a reinforcement learning model, an unsupervised learning model, a supervised learning model, a random forest model, text to vector model, and a transformer model.

16. The apparatus of claim 13, the processor further: extracting a property location from the property query using a location model ; and translating the property location into geocoded coordinates.

17. The apparatus of claim 16, wherein the location model is trained on pairs of user queries and property locations.

18. The apparatus of claim 13, wherein the privacy component grants access to a given user for the given user’s property information and public property information.

19. The apparatus of claim 13, wherein the response information comprises at least one of a property map and an itemized list.

20. A computer program product comprising a non-transitory computer readable storage medium storing code executable by a processor to perform: extracting description text from property information using a text extractor; identifying property information from the description text using a process model, wherein the property information comprises a rent and an area and the process model is trained on a training data set of lease revenue, rentable square feet, usable square feet, and / or vacant square feet; organizing response information from the property information to conform to a normalized rent table; and presentingthe response information.STATEMENT UNDER ARTICLE 19 (1 )The amendments claim the elements embodied in Figure 6 in the independent claims and move the elements embodied in Figure 5Ato dependent claims. The amendments to not impact the description and the drawings.

Images