Investor FAQs

Artificial Intelligence allows for large or complex bodies of data to be analyzed quickly and more effectively than humans can, due to superior processing speed and scalability.

Today, AI extends our own cognitive capabilities and lowers the cost and time it takes to gain insights from large data sets. In the future, AI will perform, and potentially out-perform, some human cognitive functions altogether. Artificial Intelligence is derived by the use of Cognitive Engines.

A Cognitive Engine is a specialized machine learning algorithm trained to do certain things, such as transcribe spoken language from audio files or recognize faces and objects in video files. No one engine can do everything well. The more specialized the engine, the higher the level of accuracy.

A cognitive engine’s task needs to be defined narrowly and then trained, in order for it to produce accurate and effective results. For example, notice that while Siri is effective in handling daily language, it doesn’t support a strong medical vocabulary. The wider the scope of needs, the lower the accuracy of results. An engine can be trained to recognize spoken English speech, but that same engine would be incapable of recognizing French or industry jargon in the case of the medical field.

Over 10 thousand engines have been developed and that number is growing exponentially.

Cognitive engines alone are not the answer to effective AI. An upper layer of AI is necessary to conduct, orchestrate and learn which cognitive engines are best to be employed for a particular task to yield the most accurate result.

For most tasks, no single engine is capable of producing the required analysis at a sufficient level of accuracy.

For example:

• A law enforcement user may need to search surveillance footage for both faces and tattoos.
• A conversation being transcribed may shift between English and Spanish.
• A user may want to search for a specific individual speaking certain words.

In each example above, multiple cognitive engines are needed to deliver the required answers for each situation. Typically, multiple cognitive engines are often needed, across multiple classes or categories of cognition to provide differing outputs that can be combined to reach the most accurate result. Applications can then be introduced to customize the output for the use case. To solve this technical problem, a Cognitive Operating System is needed.

A computer’s operating system connects a processor, peripherals and applications so they can communicate, working together to display what you see on your screen and respond to your commands. A cognitive operating system works in a similar fashion. It provides a common software infrastructure that lets you use end-to-end, AI-powered solutions — from data ingestion to intelligent data analysis — in either general or industry-specific applications. With thousands of limited cognitive engines on the market, you need an operating system to manage and orchestrate them all, so you can use them with general or industry-specific applications to solve real-world challenges.

aiWARE is Veritone’s cognitive operating system. What sets aiWARE apart is Veritone’s proprietary layer of AI that sits atop all the Cognitive Engines in the Operating System itself. This layer, called Conductor, is used to intelligently orchestrate multiple engines simultaneously. Machine learning is leveraged to record the strengths and weaknesses of each engine in real-time, delivering the highest value and most accurate results.

This approach delivers superior analysis of the provided data, outperforming the results any one engine could otherwise supply.

The benefits of an operating system for AI

For Business Users: An operating system for AI makes it possible to quickly and accurately analyze massive amounts of diverse information — audio, video and text — from multiple sources and formats in a faster, more scalable and more cost-effective way than humans can.

For IT Decision Makers: With the pace of technological development, implementing any single technology is perilous because it eventually becomes outdated. Most single point solutions also can not be expanded to add additional applications or cognitive capabilities. Our operating system for AI combats that problem by allowing you to add best-in-class applications and cognitive engines from our broad and constantly growing ecosystem, essentially future-proofing your technology investment.