• Rob Brown

COVID-19 rebound ideas

The COVID-19 virus has dealt the economy quite a blow, but a rebound is coming and holds the promise of doing things better than before.


Image credit: The Economist

Talk of a post-COVID economy is happening all over the world. While economic recovery may be tempered by caution, there is no lack of motivation to get back to business quickly.


As with any major disruption, there will be changes. The new economy will not be identical to what we had before and therein lies the potential for new opportunities. With all the recent advances in technology, what can new computing power mean for you as you participate in a global economic rewiring?


This is where the No-Code Toolkit in nSymbol Tag fits in. It puts new computing power in your hands and makes new interactions possible that just a few years ago could not exist. This post reviews the major capabilities of the toolkit and suggests a few ways that they can empower you in the post-COVID economy.


Blockchain

There are many ways that blockchain technology can make your working life easier. Lots of people are thinking and writing about these areas because collectively they offer a profound improvement in how we identify ourselves and conduct business transactions, both online and offline.


One of the most significant blockchain features is person to person transactions. You can securely exchange money for products or services without using a bank or transfer service as an intermediary. This is discussed further in the related Smart Contract and Bill of Sale blog posts.


Another advantage of person to person transactions is the contactless exchange of money. Science has proven that COVID-19 and other viruses can be transmitted via cash. When funds are exchanged via smart contract there is no paper money involved, so this risk is easily eliminated,


Related to the exchange of funds is near field communication (NFC). NFC refers to the ability of smart phones to detect one another over small distances (<= 10 cm) using wireless connectivity.


Another advantage of blockchain technology is the general concept of notary services. These are discussed here in the most general sense, where a notary service acts as an official, impartial witness for possessing or signing a variety of documents. For example, in some cases a simple image of a certificate with your business name and address will be enough to assure another party that you are qualified to provide a certain product or service. In other cases, using an official notary public or commissioner will be the preferred option.


This is one area of rapid change. Blockchain networks can easily prove the existence of a document at the time of upload and rigorously track modifications over time. While there are legal and security implications to using this approach, government and other agencies all over the world are starting to think about how blockchain can be used to add value and reduce cost during common interactions.


Taking this idea one step further is the emerging concept of Self-Sovereign Identity (SSI). SSI puts control of identity and personal information firmly in the hands of individuals. Emerging leaders in this area like the Sovrin Network are using blockchain technology to make SSI a reality.


A simple example of SSI can explain the main concepts. If an adult enters a lounge they must be prepared to prove their age. This is often done using a drivers license which unfortunately also provides name, birthdate, address and other personal information. If SSI is used instead, proving your age can be handled quite differently:


  1. Select a trust anchor who can confirm your age (e.g. government, bank, university)

  2. Request an attestation that confirms you are the required drinking age or older - this is different than providing a birthdate which includes more personal information than is absolutely necessary

  3. Store the attestation on a mobile device

  4. Share the attestation whenever needed


If you lose your phone or buy a new one, simply request new attestations from the appropriate trust anchors (presumably at no cost). No central authority needs to store your attestations and you only share what you need to with other individuals or organizations.


Attestations are digitally signed to guarantee they are valid and trustworthy. They can be used to prove citizenship, education, business licenses or potentially the presence of COVID-19 antibodies (if that becomes a useful thing to prove during the rebound).


It should be noted that even though the Sovrin Network is designed for Hyperledger blockchains, efforts are underway to provide similar functionality on the Ethereum network. While general purpose SSI is not yet part of Tag, we welcome opportunities to leverage it for private or public blockchain networks.


Machine learning

Along with blockchain, machine learning (ML) has been making major leaps in recent years. ML models are now routinely performing tasks that humans find challenging or tedious. They make it easy to imagine a much smarter collective future.


One of ways ML has been held back is access to tools. Most people don't have easy access to ML models and building them from scratch requires specialized training.


Tag incorporates several powerful ML models created by Amazon Web Services. While there is a small per-use fee for calling these models, the value they offer when doing business in a new way can be much larger in comparison.


Building on some of the blockchain examples above, adding automated translation or text-to-speech functionality can expand the potential market for smart contracts. For example, legal text for a Bill of Sale contract could be translated to a different language on demand. The translated text may be a bit formal sounding, but should get the job done in most cases. Similarly, using text-to-speech can make legal text and other content accessible to the visually impaired.


Speech-to-text models can be used to provide a text transcription for voice recordings. This could be used to sign a contract verbally, trigger contract events or add additional terms that are mutually agreeable. It could also be used to gather content of any kind verbally in a "hands-free" mode.


Digitization refers to the process of converting visual information into a digital asset. When the input is a PDF file or image (perhaps taken with a mobile device), the text extraction service can be used. It uses machine learning to visually analyze the input and return paragraphs that contain the detected content. This could be used to scan receipts and detect product items and prices for a delivery service.


There are several services built around comprehension. They accept one or more files containing content and return insights including detected sentiment, key phrases, dominant language, entities and topic models. These could be used to analyze emails or social media comments to detect hot topics or strong opinions.


Comprehension is also useful for pulling information out of files. When looking at folders full of files, you can detect the most commonly occurring tags (terms) and view links between all files that share those terms. When looking at files, you can detect significant entities including people, organizations, dates, locations and more. This can be used to convert entities into live data hooks (or other dynamic text) using content automation functionality which is discussed next.


Content automation

Communication is an essential part of doing business. This includes customer content (e.g. invoices, support information, emails, marketing and more) or documentation that must be shared with regulatory agencies and other stakeholders.


Content automation can reduce the time it takes to create customer content and improve quality. Rather than saving a copy of something as a Word document and then manually editing for new customers, take a bit more time and organize it into callable templates. Never again forget to change a client name or pick up a he/she copy/paste error.


There are many ways this approach can plug into a business process. For example:


  • Automate reports or other documents that are created on a recurring basis

  • Create up-to-date marketing content that merges product information from a CSV file

  • Supply chain events could be recorded (on blockchain or in a database) and used to generate a summary of all product transfers during delivery

  • Personalize customer emails using data pulled from other systems

  • Blockchain smart contracts can generate legal text to reflect parties and terms

  • The list goes on...


The process is quite simple. First, the data to be used is defined and templates created to merge data with formatted text. Next, data is gathered using auto-generated forms or by pulling data from other systems. Finally, the desired content is generated for use in Word documents, emails, website content or other purposes.


In situations where multiple files are used as inputs, task pipeline instructions can be used to do more work. For example, you could generate a unique output document for every file in an input folder using if/then and other logic instructions to customize the output. This could be used to generate an invoice for every file in a folder that tracks deliveries or product assembly.


It is also possible to generate vector graphics using content automation. This is a specialized type of content automation that supports graphics-specific instructions like circle, rectangle, line and fill. An example use case is to generate a diagram of inventory bins showing current levels of inventory (read from a database or service) as fill lines in the diagram and %-full values. This kind of visual communication can be very effective and lends itself well to automation.


Knowledge graphs

While most people might think that they never use knowledge graphs, quite the opposite is true. Search engines use them to categorize your search terms and provide the best results (and most effective ads) possible. Many online stores use them to categorize products or services as do an increasing number of health and wellness services.


The reason for their popularity is that graphs are a great way to organize knowledge. They not only define useful terms, but also link to other related terms that together describe a much richer context.


There are many public knowledge graphs (also called ontologies, taxonomies or semantic models) that contain information of value for specific business segments. These can be queried for things like billing codes, product categories and more. For example:


  • Finance: FIBO (things of interest in financial applications), FRO (regulatory compliance), ACTUS (interoperable digital financial assets)

  • Medical: FHIR (electronic health records), ICD-10-CM (International Classification of Diseases), RxNorm (clinical drugs), SNOMED (clinical terminology), OGMS (Ontology for General Medical Science), MedDRA (data entry, retrieval, analysis, and display), CPT (Current Procedural Terminology)

  • Sector-specific: LKIF (Commonsense-based Legal Ontology), IRO (Insurance Regulation Ontology)

  • Other: PROV (provenance metadata), Good Relations (e-commerce), schema.org (web search)


Knowledge graphs also provide an excellent way to reconcile similar systems or data. For example, after an acquisition a company may have multiple ways to identify the same thing. A part in one system can be called a component in another system. Knowledge graphs can state that System 1 Part is the same thing as System 2 Component when both are loaded into a graph together, which lets queries use either term interchangeably and provides a powerful unified view of both systems.


Regular computer users don't have an easy way to create knowledge graphs because that task typically requires specialized training. One goal of the No-Code Toolkit is to change that.


An earlier version of our software included a full ontology editor (based on Apache Jena) similar to Protege. While this certainly was powerful, we found that regular users found it too complex, and that it contained more logic-modeling functionality than they needed or wanted to know about.


Instead, we are planning an interface that is better described as a smart taxonomy editor. Using this editor, users create terms that may inherit from other terms (subclasses). For example, the term Car could inherit from Vehicle to indicate a more specialized term.


A slightly more complex "ontology mode" would allow individual occurrences of terms (instances) to store data values (properties). For example, Car could define a manufacturer property that stores an instance of the Manufacturer term (class). This is a portion of what Protege supports but ignores concepts like similar/distinct, keys or anonymous classes and properties.


We are very interested in finding other intuitive ways to allow regular users to create custom knowledge graphs. We believe this area will grow in importance over time as more individuals and organizations recognize the tremendous value that can be realized.


Note that Tag also has the ability to attach automated reasoners to knowledge graphs. This is an advanced feature that can be made available (configured for use) by request.


Summary

This post discussed several ways in which the No-Code Toolkit in nSymbol Tag can be used to improve business process in a post-COVID-19 economy. While most of the above ideas are already implemented in Tag, some are aspirational and waiting for real-world use cases.


Please contact us with suggestions for new ideas, or recommendations for markets or user groups that might benefit from our approach.


9 views

© 2020 by nSymbol Technology Inc.