Q&A with Scott Cohen of Jaxon, Inc
Jaxon is an AI platform that automates the process of labeling data, eliminating the biggest bottleneck in the AI creation process. Currently, armies of humans are manually labeling data (e.g. social media posts, call/chat logs, emails, documents, etc.) used to train AI models and they take months to produce enough. Jaxon synthetically labels data in minutes and curates it along the way, producing better training and test sets that improve model accuracy.
Join Jaxon CEO Scott Cohen for a private Family Office Insights Webinar/Teleconference on Jaxon’s Training Data Platform (TDP).
June 9, 2021 at 2:15pm-3:15pm EST
Family Office Insights is a voluntary, “opt-in” collaborative peer-to-peer community of single family offices, qualified investors and institutional investors. Join the community here www.familyofficeinsights.com
What problem is Jaxon solving?
Artificial intelligence is currently powered by armies of humans. Machines cannot ‘read’ natural language like humans do and require labeled data to ‘learn’ (adjust neural network weights). Massive efforts are underway to manually label data; typically hundreds of thousands of examples are needed to train machine learning models, and even more for deep learning. Human-powered approaches are extremely slow (take months), expensive (projects usually start in the 6 figures), wildly inconsistent, biased, and error-prone (often 3+ humans are put on the same labeling task to increase accuracy).
Who needs Jaxon?
Jaxon is a powerful technology enabler for any organization that is applying artificial intelligence to text. With push-button functionality, non-data scientists can use Jaxon to create custom classifiers on demand, opening up a wide range of opportunities.
Jaxon has established a foothold into the Consumer space, having won a Top 10 Retailer as an early customer last year. Every retailer wants to understand their customers better and predict intent (increasing conversions and retaining loyalty). Customer service call/chat
transcripts, reviews, surveys, and social media produce valuable content that data science teams struggle to properly process. The biggest bottleneck is the amount of labeled data needed to properly train machine learning models. Jaxon saves inordinate amounts
of time and produces better results.
Jaxon is currently expanding into the Financial Services and Insurance verticals. Jaxon has also won 3 awards with the US Air Force and has a huge opportunity across the Department of Defense.
What we are doing with Jaxon is starting to crack the questions “What does it mean when we use these words in sentences?” and “How are sentences related to one another?” and “Is this word close to this other word and why? - context that is easy for humans, but nearly impossible for machines without the guidance of labels.
Jaxon is a general purpose platform, and is applicable to a wide range of use cases. Data Science teams are buried in requests for custom models coming from across the organization; Customer Service, Marketing, Sales, HR, etc. In the interest of focus and traction,
the Jaxon team has been focusing on marketing a use case from Retail that transcends verticals, Customer Service Call Transcript Analytics.
Another application is for investment analysts - taking all of the data they have to sift through (sell-side reports, news, industry journals, social media, etc.) and classifying and then relevance ranking them. Problem specifications are defined by the users, so
they can both optimize time and ensure high trust in their recommendations. The key value is that they create custom AI models that are incorporated into their existing workflows.
Scott Cohen of Jaxon, Inc
Scott is a serial technologist and patented inventor with a penchant for pushing the envelope of innovation. Scott was a pioneer in the wireless imaging industry, having created one of the first systems able to send images and data to mobile devices for Federal, State, and Local First Responders. This system grew into an interoperability platform for seamless cross-domain data exchange for civilian and defense agencies, healthcare, and private-sector stakeholders. Through his company, DropFire, Scott successfully licensed and commercialized intellectual property from MIT to revolutionize the data management domain. With a mission of “Better Data, Better Insights”, DropFire’s data transformation platform enabled Big Data teams to collaborate and drive analytics with a wider variety of higher-quality data. DropFire pushed the bounds of machine reasoning applied to automatic data transformation. As Co-Founder, Scott then brought his passion for cutting-edge technology to BigR.io, a Big Data and Machine Learning consulting firm that specializes in building high-volume, highly-available systems, deep neural networks, and custom advanced analytics portals for the Fortune 500. Scott holds a Bachelor of Science from Union College, an MBA from Northeastern University, and a Master of Science from the Fulton School of Engineering at Arizona State University with coursework performed at the Sloan School of Management at the Massachusetts Institute of Technology (MIT).