Skip to main content

Reveal Review Publication

Reveal AI Model Library

Reveal AI comes with a growing number of pre-trained starter models to facilitate analysis of storybook data. These models are ready out of the box to be applied to your data. The list of models is augmented as new versions of Reveal AI are released.

There are two types of models in the AI Model library: COSMIC models that identify particular types of content and custom entity models that identify types of entities within sentences such as organizations or product names. The tables below set out the current library listing with version and description.

In addition, COSMIC models can be created with COSMIC enhancements: models can be created using mBERT (multilingual deep learning technology) or using High Precision Classification (HPC).

For more information on MBERT, see MBERT Explained. For more information on HPC, see High Precision Classification.

Table 1. AI Models

Name

COSMIC Version

MBERT Version

HPC Version

Description

Privileged Content

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies documents that contain privileged content. This includes conversations involving requests for legal advice, legal advice itself, and preparing documents for depositions.

Asking for Advice

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications in which the participants solicit or share advice. When used with other models, this model adds the context of advice with other behaviors of interest, like privileged content or career goals.

Advertisements & Promotions

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications that can include advertisements, newsletters, and other forms of promotional material. This model may be used to filter out irrelevant documents to reduce noise within the data set.

Gifts & Entertainment Kickbacks

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications that discuss gifts and forms of entertainment that may have a monetary value associated with it, i.e. sporting events, theater performances, concerts. This model may be used in connection with domain filters to detect instances of bribery & kickbacks.

Document Management

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications that discuss modifying or updating documents. This model can be used with Reveal AI’s Privileged Content model.

Contracts

3.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies contractual agreements in the form of loose files, attached documents, or cut-and-pasted excerpts of contracts embedded in emails.

Sexually Explicit Comments

1.1, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications that include descriptive language related to sexual acts or inappropriate behavior. These communications may be general in nature or targeted at a specific person or situation.

Comments on Appearance

1.1, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications related to a person's attire or physical attributes. These communications may have positive or negative connotations, and may exist in the form of jokes, compliments, or advice.

Employment & Career Advancement

1.1, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications about promotions, job opportunities, and workplace performance.

Bullying & Toxic Behavior

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies aggressive language with the intent to cause someone mental harm or shame.

Hate & Discrimination

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications of animosity or disparagement toward an individual or a group due to their race, color, national origin, sex, disability, religion, or sexual orientation.

Insulting & Vulgar Behavior

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communication having offensive language.

Threatening Behavior

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications with offensive language and strong intent to inflict physical or mental harm to someone else.

Pricing & Fees

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications about pricing or fees for a good or service with colleagues, clients or outside vendors.

Out of Office

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies automatically generated out of office notifications.

Sports News

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications discussing news about various professional sports.

Personal & Family Events

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications related to family and personal plans outside of work.

Work Events

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

--

Identifies communications related to social activities with work colleagues, such as team outings, dinner plans and happy hours.

Business Meetings

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

1.0, compatible with Reveal AI 3.00 & above

Identifies communications with mentions of business meetings, such as business dinners, business office visits, and mentions that imply that the participants met or will meet. This includes meetings on collaboration platforms.

Trading Conversations

1.0, compatible with Reveal AI 2.80 & above

1.0, compatible with Reveal AI 2.90 & above

1.0, compatible with Reveal AI 3.00 & above

Identifies communications related to trading, stock market, stock prices, and investments.

Table 2. Custom Entity Models

Name

COSMIC Version

Description

Street Address

1.0, compatible with Reveal AI 2.80 & above

Identifies street addresses within conversations as an entity.

Phone Number

1.0, compatible with Reveal AI 2.80 & above

Identifies local and toll-free phone numbers within conversations as an entity, including situations where a phone number includes letters combined with numbers, like 1-800-FLOWERS.