ai-one analyst toolbox introduction march 2016

13
The Analyst Toolbox Enterprise A.I. Tom Marsh, COO March 2016 www.analyst-toolbox.com

Upload: boulder-equity-analytics

Post on 28-Jan-2018

242 views

Category:

Software


2 download

TRANSCRIPT

Page 1: ai-one Analyst Toolbox Introduction March 2016

The Analyst Toolbox Enterprise A.I.

Tom Marsh, COO March 2016 www.analyst-toolbox.com

Page 2: ai-one Analyst Toolbox Introduction March 2016

ai-one Core Technology: With Nathan we create intelligent agents and applications.

“Agents are autonomous software that achieve goals by learning, adapting and reacting to the environment.”

We use them in two ways: • Model the environment with Nathan

using unstructured data. • Model the user with Nathan’s ability

to learn the user preferences.

Page 3: ai-one Analyst Toolbox Introduction March 2016

Agent learns every use of every word for the idea(s) a user trains it for… Nathan’s neural model is a dynamic fingerprint.

Detect Ideas, not Keywords using Fingerprints

Humans combine words to form ideas... Nathan converts them into static fingerprints

Applications compare fingerprints to find, score, classify similar ideas to the one he learned.

Documents, email chat, social media

Idea Agent Fingerprints

Page 4: ai-one Analyst Toolbox Introduction March 2016

Idea Detection: Score and Classify Solution

Content Management

System – Data Warehouse

Enriched Data scored and classified by idea(s)

Source Content

Page 5: ai-one Analyst Toolbox Introduction March 2016

Analyst Toolbox – Solution Workflow

Uses

• Research

• Knowledge Management

• Compliance

• Survey Coding

Page 6: ai-one Analyst Toolbox Introduction March 2016

The Analyst Toolbox

BrainDocsICE (released) - uses intelligent agents for finding, classifying and organizing content by concept “idea” BrainBrowser (Fall 2016) – enables users to analyze a document and “find something like this” on the web to find & build sources BrainView (Fall 2016) – visualize learned associations in content (sources: social media, documents, user comments and review notes) to explore patterns and sentiment

Page 7: ai-one Analyst Toolbox Introduction March 2016

BrainDocs– Concept Search and Classification

Page 8: ai-one Analyst Toolbox Introduction March 2016

BrainBrowser– Concept (re)Search for the Web

Page 9: ai-one Analyst Toolbox Introduction March 2016

BrainView – Discovery, sentiment and navigation

Page 10: ai-one Analyst Toolbox Introduction March 2016

Product and Use Cases

BrainDocs Application • Cloud or on-premise • API to automate workflows • Integrates with Tableau and

other BI tools • BrainDocs ICE available now

Technology: • Nathan API for language • Fast and scalable • Runs on standard PC class VM

Solutions • Compliance and Audit Tool • Qualitative Portfolio Metrix • Strategic Planning Support • Classification of free text in operations

SQL databases • Marketing Survey coding • Competitive Intelligence • Research, search and curation

Page 11: ai-one Analyst Toolbox Introduction March 2016

Customer Feedback

Aerospace company completed a detailed technical evaluation of BrainDocs (prototype version). Some excerpts from the report: • "In manual reviewing of the results and given the rankings it was

determined that the BrainDocs engine was very accurate with its analysis."

• "The “brain” itself seems to work great and is able to order documents by their relevance. "

• "The evaluation of BrainDocs found that the technology is very capable of its core feature (concept based text interpretation and search)."

• Regarding our color coding of relevance: “It was found that, documents placed in the green category contained concepts almost identical to what was described in the agent. Documents placed in the yellow category contained concepts relevant to the agent, such as the idea of a plane crash, death and injury, the date, and location. Documents placed in the gray category may have had similar keywords, such as "plane" and “airline” but did not contain the same developed concepts as those in the green and yellow categories.”

Page 12: ai-one Analyst Toolbox Introduction March 2016

ai-one | Quick Facts

Mission: Embed intelligence in every computing device.

Business model: Technology licensing, consulting and development

Quick Facts

• Founded 2003 as R&D company to commercialize discovery in neural -biologically inspired computing

• US parent C-Corporation established 2009, private • Offices in Berlin, Zurich, La Jolla (San Diego, CA) • 9 staff on core technology team; 34 staff in joint ventures • Customers and partners in 14 countries around the world.

Page 13: ai-one Analyst Toolbox Introduction March 2016

Thank you. Tom Marsh, COO

ai-one inc. 5666 #104 La Jolla Blvd. La Jolla, CA 92037 Ph: +18585310674 [email protected]

Follow us on Twitter @ai_one Website www.ai-one.com

www.analyst-toolbox.com